[go: up one dir, main page]
More Web Proxy on the site http://driver.im/

WO2021232775A1 - Video processing method and apparatus, and electronic device and storage medium - Google Patents

Video processing method and apparatus, and electronic device and storage medium Download PDF

Info

Publication number
WO2021232775A1
WO2021232775A1 PCT/CN2020/137690 CN2020137690W WO2021232775A1 WO 2021232775 A1 WO2021232775 A1 WO 2021232775A1 CN 2020137690 W CN2020137690 W CN 2020137690W WO 2021232775 A1 WO2021232775 A1 WO 2021232775A1
Authority
WO
WIPO (PCT)
Prior art keywords
target object
video
target
learning
detection
Prior art date
Application number
PCT/CN2020/137690
Other languages
French (fr)
Chinese (zh)
Inventor
孙贺然
王磊
白登峰
夏建明
曹军
Original Assignee
北京市商汤科技开发有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 北京市商汤科技开发有限公司 filed Critical 北京市商汤科技开发有限公司
Priority to KR1020217021262A priority Critical patent/KR20210144658A/en
Priority to JP2021538705A priority patent/JP2022537475A/en
Publication of WO2021232775A1 publication Critical patent/WO2021232775A1/en

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/50—Context or environment of the image
    • G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/10—Services
    • G06Q50/20—Education
    • G06Q50/205—Education administration or guidance
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/40—Scenes; Scene-specific elements in video content
    • G06V20/41—Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/20—Movements or behaviour, e.g. gesture recognition

Definitions

  • the present disclosure relates to the field of computer vision, and in particular to a video processing method and device, electronic equipment, and storage medium.
  • the present disclosure proposes a video processing solution.
  • a video processing method including:
  • the video wherein at least part of the video frames in the video contain the target object; according to the video, detect at least one type of learning behavior of the target object during the course of watching the teaching course; when the target object is detected
  • the learning state information is generated according to at least a part of the video frame containing the at least one type of learning behavior and/or the duration of the target object performing the at least one type of learning behavior.
  • a video processing device including:
  • a video acquisition module configured to acquire a video, wherein at least part of the video frames in the video contain the target object
  • the detection module is configured to detect at least one type of learning behavior of the target object in the process of watching the teaching course according to the video;
  • a generating module configured to perform the at least one type of learning according to at least part of the video frame containing the at least one type of learning behavior and/or the target object in the case of detecting that the target object performs at least one type of learning behavior The duration of the behavior to generate learning status information.
  • an electronic device including:
  • a processor a memory for storing executable instructions of the processor; wherein the processor is configured to execute the above-mentioned video processing method.
  • a computer-readable storage medium having computer program instructions stored thereon, and when the computer program instructions are executed by a processor, the foregoing video processing method is implemented.
  • a computer program including computer readable code, and when the computer readable code is executed in an electronic device, a processor in the electronic device executes the method for implementing the video processing method described above. .
  • the video frame containing the learning behavior when it is detected that the target object has at least one type of learning behavior, can be used to generate intuitive learning state information, and the quantified learning can be generated according to the duration of the learning behavior Status information.
  • the above-mentioned methods can be used to flexibly obtain learning status information with evaluation value, which is convenient for teachers or parents and other relevant personnel and institutions to effectively and accurately grasp the learning status of students.
  • Fig. 1 shows a flowchart of a video processing method according to an embodiment of the present disclosure.
  • Fig. 2 shows a block diagram of a video processing device according to an embodiment of the present disclosure.
  • Fig. 3 shows a schematic diagram of an application example according to the present disclosure.
  • Fig. 4 shows a block diagram of an electronic device according to an embodiment of the present disclosure.
  • Fig. 5 shows a block diagram of an electronic device according to an embodiment of the present disclosure.
  • Fig. 1 shows a flowchart of a video processing method according to an embodiment of the present disclosure.
  • the method can be applied to a video processing device, and the video processing device can be a terminal device, a server, or other processing equipment.
  • terminal devices can be User Equipment (UE), mobile devices, user terminals, terminals, cellular phones, cordless phones, personal digital assistants (Personal Digital Assistants, PDAs), handheld devices, computing devices, vehicle-mounted devices, and mobile devices.
  • UE User Equipment
  • PDAs Personal Digital Assistants
  • the data processing method can be applied to a cloud server or a local server
  • the cloud server can be a public cloud server or a private cloud server, which can be flexibly selected according to actual conditions.
  • the video processing method can also be implemented by a processor invoking computer-readable instructions stored in the memory.
  • the video processing method may include:
  • Step S11 Obtain a video, where at least part of the video frames in the video contain the target object.
  • Step S12 according to the video, detect at least one type of learning behavior of the target object in the process of watching the teaching course.
  • Step S13 in a case where it is detected that the target object performs at least one type of learning behavior, generate learning state information according to at least part of the video frames containing at least one type of learning behavior and/or the duration of the target object performing at least one type of learning behavior.
  • the target object can be any object whose learning status information is acquired, that is, an object with learning status evaluation requirements, and its specific implementation form can be flexibly determined according to actual conditions.
  • the target object may be students, such as elementary school students, middle school students or college students, etc.; in a possible realization method, the target object may be adults with advanced studies, such as adults participating in vocational education and training. Or the elderly who study in senior colleges, etc.
  • the video may be a video recorded by the target object while watching the teaching course.
  • the realization form of the teaching course is not limited. It may be a pre-recorded course video, or a live course or a teacher’s site. Courses taught, etc.; at least part of the video frames in the video can contain the target object, that is, the appearance of the target object in the recorded video can be flexibly determined according to the actual situation.
  • the target object may always be in the video.
  • the target object may also not appear in the video frame at certain moments or certain periods of time.
  • this scene can be an online scene, that is, the target object watches the teaching course through online education methods such as online classrooms;
  • this scene can also be an offline scene, that is, the target object can watch the teaching course taught by the teacher on the spot through the traditional face-to-face teaching method, or the target object can watch the video through the classroom and other specific teaching places. Or teaching courses played in other multimedia formats.
  • the video can be a real-time video, such as a video recorded in real time by the target object in the online classroom learning process, or the target object is captured by a camera deployed in the classroom while the target object is in the classroom.
  • the video can also be a recorded video. For example, after the target object learns through an online classroom, the recorded playback video of the target object learning, or the target object after the classroom has a class, Complete classroom learning videos collected through cameras deployed in the classroom.
  • the subsequent disclosed embodiments all take the video recorded in real time during the online classroom learning process of the target object as an example to illustrate the video processing process.
  • the video processing process in other application scenarios can be flexibly extended with reference to the subsequent disclosed embodiments, which will not be repeated here.
  • step S12 can be used to detect at least one type of learning behavior of the target object in the process of watching the teaching course.
  • the type and quantity of the detected learning behaviors can be flexibly determined according to actual conditions, and are not limited to the following disclosed embodiments.
  • the learning behavior performed by the target object may include at least one of the following behaviors: for example, performing at least one target gesture, expressing the target emotion, paying attention to the display area of the teaching course, and generating at least one behavior with other objects.
  • This kind of interactive behavior not appearing in at least part of the video frame in the video, closing eyes, and eye contact in the display area of the teaching course.
  • the target gesture can reflect certain preset gestures that the target object may produce during the course of watching the teaching course.
  • the specific implementation form can be flexibly set according to the actual situation. For details, please refer to the subsequent disclosed embodiments. Do unfold.
  • the target emotion can be some emotions of the target object that reflect the true feelings of the teaching course during the process of watching the teaching course.
  • the specific realization form can also be flexibly set according to the actual situation, and will not be expanded here.
  • the display area can be the display area of the teaching course video in the online classroom.
  • the display area can be these terminals
  • At least one kind of interaction behavior with other objects can be the learning-related interaction generated by the target object and other objects related to the teaching course during the course of watching the teaching course.
  • the realization form of other objects can be flexibly determined according to the actual situation.
  • other objects can be teaching objects, such as teachers, etc.
  • other objects can also be learning objects other than the target object in the teaching process, such as the target object’s Classmates, etc.; the interaction behavior with other objects can be flexibly changed according to the different objects.
  • the interaction with other objects can include receiving sent by the teacher For example, receiving small red flowers from the teacher or commendation by name, etc.
  • the interaction with other objects can include answering the teacher’s questions or according to
  • the interaction with other objects can include group mutual assistance, group discussion, or group study.
  • At least part of the video frame does not appear in the video. It may be that the learning object has left the teaching course at certain moments or certain time periods. For example, the target object may temporarily leave the current online due to personal reasons during the online learning process. Learning equipment, or leaving the shooting range of the current online learning equipment, etc.
  • Eye-closing can be the closed-eye operation performed by the target object in the process of watching the teaching course.
  • the eye contact in the display area of the teaching course can be the display area of watching the teaching course.
  • the situation of eye contact in the display area of the teaching course can further determine the situation that the target object has not watched the display area of the teaching course, etc.
  • comprehensive and flexible behavior detection can be performed on the learning process of the target object, thereby improving the comprehensiveness and accuracy of the learning status information obtained according to the detection, and making it more flexible and accurate. Grasp the learning status of the target object.
  • step S12 which type or types of detections are performed on the various learning behaviors in the above disclosed embodiment can be flexibly set according to actual conditions.
  • the various learning behaviors mentioned in the above disclosed embodiments can be detected at the same time, and the specific detection methods and processes can be detailed in the following disclosed embodiments, which will not be expanded here.
  • the learning state information may be generated according to a video frame containing at least part of the at least one type of learning behavior and/or the duration of the target object performing at least one type of learning behavior.
  • the specific implementation form of the learning status information can be flexibly determined according to the type of learning behavior and the corresponding operation performed.
  • the learning state information in a case where the learning state information is generated based on video frames that at least partially contain at least one type of learning behavior, the learning state information may include information composed of video frames; in a possible implementation manner In the case of performing at least one type of learning behavior according to the duration of the target object, the learning state information may be data information in digital form; in a possible implementation, the learning state information may also include video frame information at the same time And data information; in a possible implementation, the learning status information can also include other status information. Specifically, how to generate the learning state information and the implementation form of the learning state information can refer to the subsequent disclosed embodiments, and will not be expanded here.
  • the video frame containing the learning behavior when it is detected that the target object has at least one type of learning behavior, can be used to generate intuitive learning state information, and the quantified learning can be generated according to the duration of the learning behavior Status information.
  • the above-mentioned methods can be used to flexibly obtain learning status information with evaluation value, which is convenient for teachers or parents and other relevant personnel and institutions to effectively and accurately grasp the learning status of students.
  • the video can be a video recorded by the target object during watching the teaching course, and the scene of the target object watching the teaching course can be flexibly determined according to the actual situation. Therefore, correspondingly, the video is obtained in step S11
  • the method can also be flexibly changed according to different scenarios.
  • the way to obtain the video may include: If the video processing device The device for online learning with the target object is the same device, and the device for online learning for the target object can be used to collect video on the process of the target object watching the teaching course; if the video processing device and the device for online learning for the target object are different devices, Then, the device for online learning of the target object can collect the video of the process of the target object watching the teaching course, and transmit it to the video processing device in real-time and/or non-real-time.
  • the way to obtain the video may include: collecting the video of the target object by deploying offline image acquisition equipment (such as ordinary cameras, shooting devices deployed in response to security requirements, etc.). Further, if the image acquisition device deployed offline can perform video processing, that is, it can be used as a video processing device, then the video acquisition process in step S11 has been completed; if the image acquisition device deployed offline cannot perform video processing, the The video collected by the offline image acquisition equipment is transmitted to the video processing device in real time and/or non-real time.
  • offline image acquisition equipment such as ordinary cameras, shooting devices deployed in response to security requirements, etc.
  • step S12 may include:
  • Step S121 Perform target object detection on the video to obtain a video frame containing the target object.
  • Step S122 Perform at least one type of learning behavior detection on the video frame containing the target object.
  • target object detection can be performed on the video to determine the video frame containing the target object in the video. After determining which video frames contain the target object, at least one type of learning behavior detection can be performed on the target object in the video frame containing the target object.
  • the method of detecting the target object can be flexibly determined according to the actual situation, and is not limited to the following embodiments.
  • the target object in the video can be detected by means such as face detection or face tracking.
  • multiple objects may be detected after the video frame is detected by means of face detection or face tracking.
  • the detected face may be further detected.
  • the image is screened, and one or more objects are selected as the target object.
  • the specific screening method can be flexibly set according to the actual situation, which is not limited in the embodiment of the present disclosure.
  • step S122 may be used to perform at least one type of learning behavior detection on the video frame containing the target object.
  • the implementation of step S122 can be flexibly changed according to different learning behaviors. For details, refer to the following disclosed embodiments, which will not be expanded here. In the case where multiple types of learning behaviors of the target object need to be detected, multiple methods can be combined to achieve multiple types of learning behavior detection at the same time.
  • the learning behavior detection of the target object in the process of watching the teaching course can be completed. That is, by performing target object detection on the video, it can be determined that this learning behavior does not appear in at least part of the video frames in the video mentioned in the above-mentioned disclosed embodiment. And further obtain the learning state information according to the video frame of the undetected target object, or calculate the time that the target object does not appear in at least some of the video frames in the video according to the video frame of the undetected target object as the learning state information .
  • a video frame containing the target object is obtained, and at least one type of learning behavior detection is performed on the video frame containing the target object.
  • the target object of the video can be used.
  • Object detection is more targeted to detect at least one type of learning behavior of the target object, thereby making the learning behavior detection more accurate, and further improving the accuracy and reliability of the subsequent learning state information.
  • step S122 can be flexibly changed according to different learning behaviors.
  • the learning behavior may include: performing at least one target gesture;
  • performing at least one type of learning behavior detection on the video frame containing the target object may include:
  • At least one of the video frames containing the target gesture is recorded as a gesture start frame
  • the number of gesture start frames and gesture end frames determine the number of times and/or time for the target object in the video to perform at least one target gesture.
  • the learning behavior detection performed on the video frame of the target object may include target gesture detection.
  • the target gesture specifically includes can be flexibly set according to actual conditions, and is not limited to the following disclosed embodiments.
  • the target gesture includes one or more of a hand-raising gesture, a thumb-up gesture, an OK gesture, and a victory gesture.
  • the target gesture can be included in the process of watching the teaching course.
  • the target object reflects the learning-related gestures according to the listening situation, such as the gesture of raising the hand used to answer the question, the content of the lecture or the teaching
  • the teacher's praise gesture tilts up, etc.
  • the OK gesture that expresses understanding or approval of the teaching content
  • the victory gesture for interaction with the instructor such as Little gesture, etc.
  • the method for detecting at least one target gesture on the video frame containing the target object can be flexibly determined according to the actual situation, and is not limited to the following disclosed embodiments.
  • the detection of target gestures can be achieved through the related algorithms of gesture recognition. For example, the key points of the hand of the target object in the video frame or the image area corresponding to the hand detection frame can be recognized, based on the hand key Gesture detection is performed on the image area corresponding to the point or hand detection frame, and based on the gesture detection result, it is determined whether the target object is performing the target gesture.
  • the detection of the target gesture can be achieved through a neural network with a gesture detection function.
  • the specific structure and implementation of the neural network with gesture detection function can be flexibly set according to the actual situation.
  • the video frame containing the target object can be Input to the neural network that can detect multiple gestures at the same time to achieve the detection of the target gesture; in a possible implementation, the video frame containing the target object can also be input to multiple neural networks with a single gesture detection function. In the network, to realize the detection of multiple target gestures.
  • the number of first thresholds can be flexibly set according to the actual situation.
  • the number of first thresholds corresponding to different target gestures can be the same or different.
  • the first threshold corresponding to the hand-raising gesture can be set to 6.
  • the first threshold corresponding to the thumbs-up gesture is set to 7. If the number of consecutive video frames containing the hand-raising gesture is detected to be not less than 6, at least one frame can be selected from the video frames containing the hand-raising gesture as the gesture.
  • the gesture start frame of the hand gesture if the number of consecutive video frames of the like gesture is not less than 7, at least one frame may be selected from the video frames containing the like gesture as the gesture start frame of the like gesture.
  • the first thresholds corresponding to different target gestures may be set to the same value. In an example, the number of the first thresholds may be set to 6.
  • the selection method of the gesture start frame can also be flexibly set according to the actual situation.
  • the first frame of the detected continuous video frames containing the target gesture can be used as the gesture start of the target gesture Frame, in a possible implementation, in order to reduce the error of gesture detection, a certain frame after the first frame of the detected continuous video frames containing the target gesture can also be used as the gesture start frame of the target gesture .
  • the gesture end frame can be determined from the video frames after the gesture start frame, that is, the end time of the target gesture in the gesture start frame can be determined.
  • the specific determination method can be flexibly selected according to the actual situation, and is not limited to the following disclosed embodiments.
  • the target will not be included. At least one of the consecutive video frames of the gesture is recorded as the gesture end frame.
  • the value of the second threshold can also be flexibly set according to actual conditions. The values of the second threshold corresponding to different target gestures can be the same or different.
  • the specific setting method can refer to the first threshold, which will not be repeated here.
  • the value of the second threshold corresponding to different target gestures can be the same, for example, it can be set to 10. That is, after the gesture start frame, it is detected that 10 consecutive frames do not contain the target gesture in the gesture start frame. It is considered that the target object has finished performing the target gesture. In this case, you can select at least one frame from the continuous video frames that do not contain the target gesture as the gesture end frame.
  • the selection method can also refer to the gesture start frame. In one example, you can select the gesture start frame.
  • the last frame in the continuous video frames is used as the gesture end frame; in an example, a frame before the last frame in the continuous video frames that do not contain the target gesture may also be used as the gesture end frame.
  • a frame before the last frame in the continuous video frames that do not contain the target gesture may also be used as the gesture end frame.
  • one or some video frames that do not contain the target object can also be set End the frame as a gesture.
  • the number of gesture start frames and gesture end frames contained in the video frame can be used to determine the number of times the target object performs a certain target gesture or certain target gestures.
  • the duration of the execution of a certain or certain target gesture, etc. The specific determination of the content related to the target gesture can be flexibly determined according to the requirement of learning state information in step S13. For details, please refer to the subsequent disclosed embodiments, which will not be expanded here.
  • the learning behavior can include: expressing the target emotion
  • performing at least one type of learning behavior detection on the video frame containing the target object may include:
  • the target object In a case where it is detected that the number of consecutive first detection frames exceeds the third threshold, it is determined that the target object generates the target emotion.
  • the target emotion can be any emotion set according to actual needs, for example, it can be a happy emotion that indicates that the target object is focused on learning, or a bored emotion that indicates that the target object is in a poor learning state.
  • the following disclosed embodiments are described by taking the target emotion as happy emotion as an example, and the case where the target emotion is other emotions can be expanded with reference to the subsequent disclosed embodiments.
  • expression detection and/or smile value detection can be used to achieve the learning behavior detection of the target object.
  • the learning behavior of expressing the target emotion can be detected only by expression detection or smile value detection.
  • expression detection and smile value detection can be used together. Determine whether the target object expresses the target emotion.
  • the subsequent disclosed embodiments are described by taking as an example the determination of whether the target object expresses the target emotion through expression detection and smile value detection. The remaining implementation manners can be expanded with reference to the subsequent disclosed embodiments, and will not be repeated here.
  • the expression detection can include the detection of the expressions displayed by the target object, for example, it can detect what kind of expression the target object displays.
  • the specific expression division can be flexibly set according to the actual situation.
  • the expression can be divided For happiness, calmness, etc.;
  • the smile value detection can include the detection of the smile intensity of the target object, for example, it can detect how big the smile of the target object is, and the result of the smile value detection can be fed back by numerical values.
  • the detection result is set to be between [0,100]. The higher the value, the higher the intensity or amplitude of the target's smile.
  • the specific expression detection and smile value detection methods can be flexibly determined according to the actual situation.
  • any method that can detect the expression or the degree of smile of the target object can be used as a corresponding detection method, and is not limited to the following disclosed embodiments.
  • the expression detection of the target object can be realized by the facial expression recognition neural network
  • the smile value detection of the target object can be realized by the smile detection neural network.
  • the structure and implementation of the facial expression recognition neural network and the smile value detection neural network are not limited in the embodiments of the present disclosure.
  • Any neural network that can realize the expression recognition function through training and the neural network that realizes the smile value detection function through training are both It can be applied to the embodiments of the present disclosure.
  • facial expression detection and smile value detection can also be realized by detecting the key points of the face and the mouth of the target object in the video.
  • the target object in the video frame is detected to show at least one first target expression, or the smile value detection result exceeds the target smile value, the target object in the video frame is considered Show the target emotion, in this case, the video frame can be used as the first detection frame.
  • the specific expression type of the first target expression can be flexibly set according to the actual situation, and is not limited to the following disclosed embodiments.
  • happiness may be used as the first target expression, that is, video frames in which the detected expression of the target object is happy may be used as the first detection frame.
  • both happy and calm can be used as the first target expression, that is, the detected expression of the target object can be a happy or calm video frame, and both can be used as the first detection frame.
  • the specific value of the target smile value can also be flexibly set according to the actual situation, and there is no specific limitation here. Therefore, in a possible implementation manner, a video frame whose smile value detection result exceeds the target smile value may also be used as the first detection frame.
  • the target in the case that a certain video frame is detected as the first detection frame, it may be determined that the target object generates the target emotion.
  • the target in order to improve the accuracy of detection and reduce the impact of detection errors on the results of learning behavior detection, the target can be determined when the number of consecutive first detection frames exceeds the third threshold.
  • the subject develops the target emotion.
  • a video frame sequence in which each frame in the continuous video frames is the first detection frame may be used as the continuous first detection frame.
  • the number of the third threshold can be a number flexibly set according to the actual situation, and its value can be the same as or different from the first threshold or the second threshold. In an example, the number of the third threshold can be 6, which means it is detected In the case where 6 consecutive frames are the first detection frame, it can be considered that the target object has the target emotion.
  • a frame from the first continuous detection frame can be selected as the target emotion start frame, and then after the target emotion start frame, the expression of the target object is not detected for 10 consecutive frames If it is the first target expression, or the smile value detection result of the target object in 10 consecutive frames does not exceed the third threshold, or the target object cannot be detected in a certain frame or a few frames, the target emotion end frame can be further determined, Then, according to the target emotion start frame or the target emotion end frame, the number and/or time of the target emotion generated by the target object are determined.
  • the specific process can refer to the corresponding process of the target gesture, which will not be repeated here.
  • the first detection frame is determined, so that when the number of consecutive first detection frames exceeds the first detection frame In the case of three thresholds, it is determined that the target object produces the target emotion.
  • the emotion of the target object in the learning process can be flexibly determined based on the expression and smile of the target object, so that the target object can be more comprehensively and accurately perceived in the learning process.
  • the emotional state in the process generates more accurate learning state information.
  • the learning behavior can include: paying attention to the display area of the teaching course;
  • performing at least one type of learning behavior detection on the video frame containing the target object may include:
  • the implementation form of the display area of the teaching course can refer to the above-mentioned disclosed embodiments, which will not be repeated here.
  • the learning behavior detection of the target object can be achieved through expression detection and face angle detection.
  • the detection of the learning behavior of paying attention to the display area of the teaching course can also be realized only by detecting the face angle.
  • Subsequent disclosed embodiments are described by using expression detection and face angle detection to determine whether the target object pays attention to the display area of the teaching course as an example. The remaining implementation methods can be expanded with reference to the subsequent disclosed embodiments, and will not be repeated here. .
  • the implementation of expression detection can refer to the above disclosed embodiments, which will not be repeated here;
  • the face angle detection can be the detection of the orientation angle of the face.
  • the specific face angle detection method can be flexibly determined according to the actual situation. Any method that can detect the face angle of the target object can be used as the face angle detection method, and is not limited to the following disclosed embodiments.
  • the face angle detection of the target object can be realized through the face angle detection neural network.
  • the structure and implementation of the face angle detection neural network are not limited in the embodiments of the present disclosure, and any neural network that can realize the face angle detection function through training can be applied to the embodiments of the present disclosure.
  • the face angle of the target object can also be determined by detecting the key points of the target object's face in the video.
  • the form of the face angle that can be detected by the face angle detection can also be flexibly determined according to the actual situation.
  • it can be determined by detecting the yaw angle and pitch angle of the target object’s face The angle of the target's face.
  • the target object is determined to focus on the display area of the teaching course, and the implementation method can be flexibly set according to the actual situation.
  • the target object of is concerned with the display area of the teaching course.
  • the video frame can be used as the second detection frame.
  • the specific expression type of the second target expression can be flexibly set according to the actual situation, and may be the same as the first target expression mentioned in the above-mentioned public embodiment, or it may be different from the first target expression mentioned in the above-mentioned public embodiment It is not limited to the following disclosed embodiments.
  • calm can be used as the second target expression, that is, the detected target object's expression is calm and the video frame whose face angle is within the range of the target face angle can be regarded as the second detection frame .
  • the face angle of the detected target object can be within the range of the target face angle, and the expression is not "other" video Frames are regarded as the second detection frame.
  • the specific range value of the target face angle range can also be flexibly set according to the actual situation, and no specific limitation is made here.
  • the target face angle range may be static.
  • the target face angle range in one example, a fixed area (such as the display screen that the target object pays attention to in an online scene) may be taken as the target face angle range when the target object views the teaching course.
  • the target face angle range can also be dynamic.
  • the target face angle range can be flexibly determined according to the current position of the teacher's movement during the lecture, that is, it can follow the teacher's movement To dynamically change the value of the target face angle range.
  • the target in the case that a certain video frame is detected as the second detection frame, it can be determined that the target object pays attention to the display area of the teaching course.
  • the target in order to improve the accuracy of detection and reduce the impact of detection errors on the results of learning behavior detection, the target can be determined when the number of consecutive second detection frames exceeds the fourth threshold.
  • the subject pays attention to the display area of the teaching course.
  • a video frame sequence in which each frame in the continuous video frames is the second detection frame may be used as the continuous second detection frame.
  • the number of the fourth threshold can be a number flexibly set according to actual conditions, and its value can be the same as or different from the first threshold, the second threshold, or the third threshold. In an example, the number of the fourth threshold can be 6. , That is, when it is detected that 6 consecutive frames are all the second detection frames, it can be considered that the target object pays attention to the display area of the teaching course.
  • the target object is not detected for 10 consecutive frames
  • the end frame of attention can be further determined , And then determine the number and/or time the target object pays attention to the teaching course display area according to the focus start frame or focus end frame.
  • the specific process can refer to the corresponding process of target gestures and target emotions, which will not be repeated here.
  • the second detection frame is determined, so that the number of consecutive second detection frames exceeds the first detection frame.
  • the target object pays attention to the display area of the teaching course.
  • the learning behavior may also include: generating at least one interaction behavior with other objects.
  • the interactive behavior reference may be made to the above disclosed embodiments, which will not be repeated here.
  • the method of detecting the interactive behavior of the video frame containing the target object can be flexibly determined according to the actual situation.
  • the interactive behavior is an online interactive behavior, such as receiving the teacher’s approval
  • the interactive behavior detection method can be directly based on the signals transmitted by other objects to determine whether the target object has an interactive behavior.
  • the method of detecting whether the target object has an interactive behavior can include: The target action of the object is recognized to determine whether the target object has an interactive behavior.
  • the target action can be flexibly set according to the actual situation of the interactive behavior.
  • the target action can include speaking after standing up or speaking with the face facing other objects. The time exceeds a certain time value, etc.
  • step S12 may include:
  • detecting the learning behavior includes: not appearing in at least part of the video frames in the video.
  • each video frame in the video may also contain video frames that do not contain the target object. Therefore, these video frames that do not contain the target object can be regarded as undetected Video frames of the target object, and in the case where the number of video frames where the target object is not detected exceeds the preset number of video frames, it is confirmed that the learning behavior of "not appearing in at least part of the video frames in the video" is detected.
  • the number of preset video frames can be flexibly set according to the actual situation.
  • the number of preset video frames can be set to 0, that is, when the video contains video frames where the target object is not detected, That is to say, it is considered that this learning behavior is not detected in at least part of the video frames in the video.
  • the preset number of video frames can also be a number greater than 0. The specific setting can be based on the actual situation. Flexible decision.
  • the learning behavior can also include closed eyes.
  • the learning behavior detection method can be closed eyes detection.
  • the specific process of closed eyes detection can be flexibly set according to the actual situation.
  • It can be realized by a neural network with closed eyes detection function.
  • it can also determine whether the target object has closed eyes or not by detecting the key points in the eyes and the eyeball. For example, after detecting the key points in the eyeball In the case of dots, it is determined that the target object has eyes open; in the case that only the key points of the eye are detected, and the key points in the eyeball are not detected, the eyes of the target object are determined to be closed.
  • the learning behavior can also include eye contact in the display area of the teaching course.
  • the learning behavior detection method can refer to the focus on the display area of the teaching course in the above disclosed embodiment.
  • the specific detection method can be flexibly changed.
  • the target object can be detected with closed eyes and face angle at the same time, and the video frame with the face angle within the target face angle range without closed eyes is used as the third detection frame. Then, when the number of third detection frames exceeds a certain set threshold, it is determined that the target object is making eye contact in the display area of the teaching course.
  • step S13 After the detection of at least one type of learning behavior of the target object is achieved through any combination of the various implementation manners of the above disclosed embodiments, it can be generated through step S13 when the target object is detected to perform at least one type of learning behavior. Learning status information.
  • the specific implementation of step S13 is not limited, and can be flexibly changed according to the actual situation of the detected learning behavior, and is not limited to the following disclosed embodiments.
  • the learning state can be generated based on a video frame containing at least one type of learning behavior. Information; or generate learning state information according to the duration of the target object performing at least one type of learning behavior; or a combination of the above two situations, both based on the video frame containing at least one type of learning behavior to generate part of the learning state information, Another type of learning state information is generated according to the duration of at least one type of learning behavior performed by the target object.
  • the learning state information can be generated based on the video frames of the learning behavior, and the learning state information can be generated based on the duration of the target object performing at least one type of learning behavior, which learning state is generated according to which type of learning behavior Information and its mapping method can be flexibly set according to the actual situation.
  • some positive learning behaviors can be corresponded to the process of generating learning state information based on the video frames containing the learning behaviors, such as performing at least one target gesture on the target object and showing a positive goal.
  • the learning state information can be generated based on the video frame containing the above learning behavior; in a possible implementation manner, it can also be Some negative learning behaviors, such as when the target object does not appear in at least part of the video frame in the video, eyes are closed, or there is no eye contact in the display area of the teaching course, can be based on the duration of the above learning behavior. Generate learning status information.
  • generating learning state information according to video frames containing at least one type of learning behavior at least in part may include:
  • Step S1311 Obtain video frames containing at least one type of learning behavior in the video as a target video frame set;
  • Step S1312 Perform face quality detection on at least one video frame in the target video frame set, and use a video frame with a face quality greater than a face quality threshold as a target video frame;
  • Step S1313 Generate learning state information according to the target video frame.
  • the video frame containing at least one type of learning behavior may be a video frame in which the target object is detected to perform at least one type of behavior in the process of learning behavior detection, such as the first detection frame mentioned in the above-mentioned disclosed embodiment, The second detection frame and the third detection frame, etc., or the video frame containing the target gesture between the gesture start frame and the gesture end frame, etc.
  • each video frame containing each type of learning behavior can be obtained according to the type of learning behavior, so as to form the target video frame set of each type of learning behavior; in a possible implementation manner, It is also possible to obtain partial frames containing each type of learning behavior according to the type of learning behavior, and then obtain the target video frame set of that type of learning behavior based on the partial frames of each type of learning behavior, which part of the frame is specifically selected, and the selection method Can be flexibly decided.
  • step S1312 may be used to select and obtain the target video frame from the target video frame set. It can be seen from step S1312 that, in a possible implementation manner, face quality detection may be performed on the video frames in the target video frame set, and then video frames with face quality greater than the face quality threshold are used as the target video frames.
  • the face quality detection method can be flexibly set according to the actual situation, and is not limited to the following disclosed embodiments.
  • the face quality can be determined by performing face recognition on the face in the video frame.
  • the completeness of the face in the video frame is used to determine the face quality; in a possible implementation, the face quality can also be determined based on the clarity of the face in the video frame; in a possible implementation, it is also
  • the face quality in the video frame can be comprehensively judged based on multiple parameters such as the completeness, clarity, and brightness of the face of the video frame; in a possible implementation, the video frame can also be input to the face quality nerve Network to obtain the face quality in the video frame.
  • the face quality neural network can be obtained by training a large number of face images containing face quality scores.
  • the specific implementation form can be flexibly selected according to the actual situation. In the embodiments of the present disclosure, No restrictions.
  • the specific value of the face quality threshold can be flexibly determined according to the actual situation, which is not limited in the embodiment of the present disclosure.
  • different face quality thresholds may be set for each type of learning behavior; in a possible implementation manner, the same face threshold may also be set for each type of learning behavior.
  • the face quality threshold can also be set to the maximum value of the face quality in the target video frame set. In this case, you can directly set the highest face quality under each type of learning behavior The video frame is used as the target video frame.
  • video frames there may be certain video frames that contain multiple types of learning behaviors at the same time.
  • the manner of processing video frames containing multiple types of learning behaviors can be flexibly changed according to actual conditions.
  • these video frames can be assigned to each type of learning behavior, and then selected from the set of video frames corresponding to each type of learning behavior in step S1312 to obtain the target video frame;
  • a video frame containing multiple types of learning behaviors at the same time can also be directly selected as the target video frame.
  • step S1313 may be used to generate learning state information according to the target video frame.
  • the implementation of step S1313 can be flexibly selected according to the actual situation. For details, please refer to the following disclosed embodiments, which will not be expanded here.
  • the video frame containing at least one type of learning behavior is obtained as the target video frame set, so that according to the target video frame set of each type of learning behavior, the video frame with higher face quality is selected As the target video frame, the learning state information is then generated according to the target video frame.
  • the generated learning status information can be based on the information obtained from the video frames with higher face quality and containing learning behaviors, with higher accuracy, so that the learning of the target object can be grasped more accurately state.
  • step S1313 can be flexibly changed.
  • step S1313 may include:
  • At least one frame of the target video frame can be directly used as the learning state information.
  • the obtained target video frame can be further selected. This selection can be random or subject to certain conditions, and then the selected target video frame is directly used as the learning state information; in one example, each target video frame obtained can also be directly equalized. As learning status information.
  • the area where the target object is located in the target video frame may be further identified, so as to generate learning state information according to the area where the target object is located.
  • the method of recognizing the target object area is not limited in the embodiment of the present disclosure. In a possible implementation manner, it can be implemented by the neural network with the target object detection function mentioned in the above-mentioned disclosed embodiment. After the area of the target object in the target video frame is determined, the target video frame can be further processed accordingly to obtain the learning state information. Among them, the processing method can be flexibly determined.
  • the image of the area where the target object is located in the target video frame can be used as the learning state information; in one example, the background outside the area where the target object is located in the target video frame can also be used as learning state information.
  • Area rendering such as adding other stickers, or adding mosaic to the background area, or replacing the image of the background area, etc., to get the learning status information that does not display the current background of the target object, so as to better protect the privacy of the target object .
  • the above method can make the final learning state information more flexible, so that According to the needs of the target object, the learning status information of the target object is more prominent, or the learning status information that protects the privacy of the target object more can be obtained.
  • Table 1 shows a learning state information generation rule according to an embodiment of the present disclosure.
  • M, N, X, Y, and Z are all positive integers, and the specific values can be set according to actual needs.
  • the parameters such as M in different rows in Table 1 may be the same or different.
  • the above-mentioned parameters such as M are only used as a schematic description, and not as a limitation to the present disclosure.
  • the amazing moment is the moment corresponding to the positive learning behavior of the target object.
  • the target object can be detected to perform target gestures such as raising hands, to generate the target emotion of happiness, or to pay attention to the display area of the teaching course, and to have a roll call with the teacher.
  • target gestures such as raising hands
  • the target emotion of happiness or to pay attention to the display area of the teaching course
  • the teacher to have a roll call with the teacher.
  • certain data processing is performed on the video, and after the data processing, further image processing is performed on the video frame to obtain the target video frame as the learning state information.
  • generating learning state information according to the duration of the target object performing at least one type of learning behavior may include:
  • Step S1321 in the case where it is detected that the time for the target object to perform at least one type of learning behavior is not less than the time threshold, record the duration of the at least one type of learning behavior;
  • step S1322 the duration corresponding to at least one type of learning behavior is used as the learning state information.
  • the time threshold can be a certain value flexibly set according to the actual situation, and the time thresholds of different types of learning behaviors can be the same or different.
  • the time for the target object to perform these learning behaviors can be counted, so as to feed back to the teacher or parent as learning status information.
  • the specific statistical conditions and the statistical time under which learning behaviors can be implemented can be flexibly set according to the actual situation.
  • the time length of these learning behaviors can be counted and used as the learning status information.
  • the duration of at least one type of learning behavior is recorded as the learning state information.
  • the video processing method proposed in the embodiment of the present disclosure may further include:
  • the segmentation method of the background area and the rendering method of the background area reference may be made to the above-mentioned disclosed embodiment for identifying the area where the target object in the target video frame is located and the rendering process after the recognition, which will not be repeated here.
  • it can be rendered by a universal template preset in the current video processing device; in one example, it can also be rendered by calling other templates in the database of the non-video processing device or Customized templates, etc. for rendering, for example, other background templates can be called from a cloud server of a non-video processing device, etc., to render the background area in the video, etc.
  • the privacy of the target object in the video can be protected, and the possibility of privacy leakage of the target object due to the lack of a suitable video capture location is reduced. On the other hand, it is also It can enhance the interest of the target object to watch the teaching course process.
  • the video processing method proposed in the embodiment of the present disclosure may further include:
  • the learning state statistical data is generated.
  • the target object contained in a video may be one or multiple.
  • the video processing method in the embodiment of the present disclosure may be used to process a single video, or it may be used to process a single video. Multiple videos are processed. Therefore, correspondingly, the learning status information of one target object can be obtained, and the learning status information of multiple target objects can also be obtained.
  • statistics can be performed on the learning state information of at least one target object to obtain a statistical result of at least one target object.
  • the statistical result may include not only the learning status information of the target object, but also other information related to the target object's viewing of the teaching course.
  • the sign-in data of the target object can also be obtained before step S12, that is, before performing learning behavior detection on the target object.
  • the check-in data of the target object may include the identity information and check-in time of the target object.
  • the specific check-in data acquisition method can be flexibly determined according to the actual check-in method of the target object, which is not limited in the embodiments of the present disclosure.
  • the learning state statistical data can be generated according to the at least one statistical result.
  • the generation method and content of the learning state statistical data can be flexibly changed according to the realization form of the statistical result.
  • the statistical result of the at least one target object is obtained by counting the learning status information of at least one target object, so as to generate the learning status statistical data according to the statistical result of the at least one target object.
  • generating the learning state statistical data according to the statistical result of at least one of the target objects includes:
  • the statistical result of the target object contained in the at least one category is obtained, and the learning status statistical data of at least one category is generated. And at least one of the devices used by the target object; and/or,
  • the category to which the target object belongs may be a category divided according to the identity of the target object.
  • the category to which the target object belongs may include at least one of the courses the target object participates in, the institution registered by the target object, and the equipment used by the target object.
  • the course that the target object participates in may be the teaching course watched by the target object mentioned in the above disclosed embodiment
  • the institution registered by the target object may be the educational institution where the target object is located, or the grade or grade of the target object.
  • the class where the target object is located, and the equipment used by the target object may be the terminal device used by the target object to participate in the online course in an online scene.
  • the statistical results of the target objects contained in at least one category can be obtained according to the category to which the target object belongs, that is, at least one statistical result of the category to which the target object belongs can be summarized to obtain the Statistics of overall learning status. For example, it can be divided according to the categories of equipment, courses, educational institutions, etc., and the statistical results of different target objects under the same equipment, the statistical results of different target objects under the same course, and the statistical results of different target objects in the same educational institution can be obtained respectively. Wait. In an example, these statistical results can also be displayed in the form of a report.
  • the statistical results of each category in the report can include not only the overall learning status information of each target object, but also the specific learning status information of each target object, such as the focus on the display area of the teaching course
  • the length of time, the length of smiling time, etc. in addition to this, it can also contain other information related to watching the teaching course, such as the check-in time of the target object, the number of check-ins, the match between the target object and the face in the preset database, Sign-in equipment and sign-in courses, etc.
  • the statistical results of at least one target object can also be visualized to obtain the statistical data of the learning state of the at least one target object.
  • the visual processing method can be flexibly determined according to the actual situation, for example, the data can be sorted into forms such as charts or videos.
  • the content contained in the learning status statistics can be flexibly determined according to the actual situation. For example, it can include the overall learning status information of the target object, the name of the teaching course watched by the target object, and the specific learning status information of the target object.
  • the actual situation is flexible.
  • the results, the number of interactions of the target object, and the emotions of the target object are organized into a visual report, and sent to the target object or other relevant personnel of the target object, such as the parents of the target object.
  • the visualized statistical data of learning status can contain text content in the form of "The subject of class is XX, the duration of concentration of A student is 30 minutes, and the concentration is concentrated, which is 10% higher than the class. % Of classmates interacted 3 times and smiled 5 times. I hereby give praise and are willing to continue to work hard" or "The subject of class is XX, B students have less concentration, and the frequency of gestures such as raising hands is lower. Parents are advised to pay close attention , Adjust the children’s study habits in time” and so on.
  • the learning state statistical data of at least one category is generated, and/or the statistical result of the at least one target object is visualized to generate the statistics of the at least one target object.
  • Statistics of learning status Through the above process, the learning state of the target object can be grasped more intuitively and comprehensively through different statistical methods.
  • Fig. 2 shows a block diagram of a video processing device according to an embodiment of the present disclosure.
  • the video processing device 20 may include:
  • the video acquisition module 21 is configured to acquire a video, where at least part of the video frames in the video contain the target object;
  • the detection module 22 is used to detect at least one type of learning behavior of the target object in the process of watching the teaching course according to the video;
  • the generating module 23 is configured to generate learning based on at least part of the video frames containing at least one type of learning behavior and/or the duration of the target object performing at least one type of learning behavior when it is detected that the target object performs at least one type of learning behavior status information.
  • the learning behavior includes at least one of the following behaviors: performing at least one target gesture, expressing the target emotion, paying attention to the display area of the teaching course, generating at least one interactive behavior with other objects, There is no eye contact, eyes closed, and eye contact in the display area of the teaching course in at least part of the video frames in.
  • the detection module is configured to: perform target object detection on the video to obtain a video frame containing the target object; and perform at least one type of learning behavior detection on the video frame containing the target object.
  • the learning behavior includes performing at least one target gesture; the detection module is further configured to: perform detection of at least one target gesture on the video frame containing the target object; When the number of continuous video frames exceeds the first threshold, record at least one of the video frames containing the target gesture as the gesture start frame; in the video frames after the gesture start frame, the continuous video frames that do not contain the target gesture When the number exceeds the second threshold, record at least one of the video frames that do not contain the target gesture as the gesture end frame; according to the number of gesture start frames and gesture end frames, determine that the target object in the video performs at least one target The number and/or time of gestures.
  • the learning behavior includes expressing the target emotion; the detection module is further used to: perform expression detection and/or smile value detection on the video frame containing the target object; in the detected video frame, the target object displays at least one When the first target expression or smile value detection result exceeds the target smile value, the detected video frame is regarded as the first detection frame; when the number of consecutive first detection frames exceeds the third threshold, it is determined The target object produces the target emotion.
  • the learning behavior includes paying attention to the display area of the teaching course; the detection module is further used to: perform expression detection and face angle detection on the video frame containing the target object; display the target object in the detected video frame In the case of at least one second target expression and the face angle is within the range of the target face angle, the detected video frame is used as the second detection frame; when the number of consecutive second detection frames exceeds the fourth threshold In this case, determine the target object to focus on the display area of the teaching course.
  • the generating module is used to: obtain video frames containing at least one type of learning behavior in the video as a target video frame set; perform face quality detection on at least one video frame in the target video frame set, The video frame whose face quality is greater than the face quality threshold is taken as the target video frame; according to the target video frame, the learning state information is generated.
  • the generating module is further configured to: use at least one frame of the target video frame as the learning state information; and/or, identify the area where the target object is located in the at least one frame of the target video frame, based on the target object In the area, the learning status information is generated.
  • the detection module is used to: perform target object detection on the video to obtain a video frame containing the target object, and use the video frame other than the video frame containing the target object as the undetected target object
  • the detected learning behavior includes: not appearing in at least part of the video frames in the video.
  • the generating module is used to record the duration of at least one type of learning behavior when it is detected that the time for the target object to perform at least one type of learning behavior is not less than a time threshold; The duration corresponding to the behavior is used as the learning status information.
  • the device is further configured to: render a background area in at least part of the video frame in the video, where the background area is an area outside the target object in the video frame.
  • the device is further configured to: collect statistics on the learning state information of at least one target object to obtain a statistical result of at least one target object; and generate statistical data of the learning state according to the statistical result of at least one target object.
  • the device is further configured to: obtain statistical results of the target objects contained in the at least one category according to the category to which the at least one target object belongs, and generate statistical data of the learning state of at least one category, wherein the target object belongs to
  • the category includes at least one of the courses the target object participates in, the institution registered by the target object, and the equipment used by the target object; and/or visualize the statistical results of at least one target object to generate the learning status of at least one target object Statistical data.
  • the functions or modules contained in the device provided in the embodiments of the present disclosure can be used to execute the methods described in the above method embodiments.
  • the functions or modules contained in the device provided in the embodiments of the present disclosure can be used to execute the methods described in the above method embodiments.
  • the application example of the present disclosure proposes a set of learning system, which can effectively grasp the learning state of students through the video processing method proposed in the above-mentioned disclosed embodiment.
  • Fig. 3 shows a schematic diagram of an application example according to the present disclosure.
  • the learning system can be composed of three parts: the user end, the educational software service (SaaS, Software-as-a-Service) backend, and the interactive classroom backend. Among them, students watch the teaching courses through the client.
  • the client can include two parts: hardware devices for learning (such as the client with Windows system or IOS system and SDK installed in the picture), and the student's login to the online classroom.
  • Application ie the user APP in the figure).
  • the education SaaS backend can be a platform built by the server of the educational institution where the student is located, and the interactive classroom backend can be a platform built by a server that aggregates data from different educational institutions and performs data maintenance, whether it is an education SaaS backend or an interactive classroom backend.
  • Data can be exchanged with the client through the API interface.
  • the process of generating learning state information may include:
  • the user terminal obtains the learning status information of each student by collecting the videos of the students watching the teaching course process and processing the collected videos.
  • the education SaaS background and the interactive classroom background call the learning status generated in different users through the API interface.
  • the user terminal processes the collected video, and the process of obtaining the learning status information of each student may include:
  • the student's wonderful moments can be obtained after the student signs in successfully, and the videos or pictures of the next wonderful moments will be uploaded to the background or the cloud. At the same time, it is also possible to choose whether the students can see the uploaded wonderful moments in real time.
  • the highlight definition rule may include: generating at least one target gesture.
  • the target gesture may include raising hand, like, gesture OK, gesture Little, etc. If a student is detected to perform the above gesture within a period of time, Then you can extract pictures or video frames from videos that contain gestures. Express the happy target emotion.
  • the smile value reaches a certain target smile value (such as 99 points)
  • a certain target smile value such as 99 points
  • the video frame performs picture or video frame extraction. Pay attention to the display area of the teaching course. If the student's face orientation is always correct within a period of time, that is, the headpose is within a certain threshold range, then pictures or video frames can be extracted from the video within this period of time.
  • the student may not be on the screen or may be unfocused, and the data can be pushed to the parents in real time through the learning situation detection, so that the parents can pay attention to the children in the first time, and correct the children’s bad learning habits in time. Auxiliary supervision.
  • the process of checking the student's academic status can be carried out after the student signs in successfully. For example, for how long in front of the camera, no one appears in front of the camera, does not watch the screen, closes eyes, etc., it is judged that the person has a low degree of concentration. In this case, it is possible to count the length of time during which the student has the above-mentioned learning behavior, and use it as the result of the academic condition detection to obtain the corresponding learning state data.
  • the specific academic condition detection configuration rules can refer to the above disclosed embodiments, which will not be repeated here.
  • learning status information including exciting moments and learning situation detection can be obtained.
  • the education SaaS backend and interactive classroom backend use API interfaces to call the learning status information generated in different client terminals to generate learning status.
  • the process of statistical data can include:
  • Report generation that is, the generation of statistical data of learning status in at least one category in the above disclosed embodiment.
  • the backend or cloud API can view student sign-in information and learning status information in different dimensions such as device, course, institution, etc.
  • the main data indicators can include: sign-in time, sign-in times, and face database (that is, the above-mentioned public The target object in the embodiment matches the face in the preset database), sign-in equipment, sign-in course, focus time, smile time, etc.
  • Analysis report (that is, the visualization process in the above disclosed embodiment generates statistics on the learning status of at least one target object).
  • the education SaaS backend or the interactive classroom backend can unify the students' performance in the online classroom into a complete academic analysis report.
  • the report explains the student’s class status through a visual graphical interface.
  • the background can also select a better situation and push it to parents or teachers, so that it can be used by institutional teachers to analyze the student’s situation and gradually assist children in improving their learning behavior.
  • the learning system can also perform background segmentation processing on the student's learning video when the student is learning through the user terminal.
  • the user terminal may provide a background segmentation function for situations where the student does not have a location background suitable for live broadcast or is unwilling to display a background image for privacy protection.
  • the SDK on the user side can support several different background templates. For example, several general templates can be preset.
  • students can also call customized templates from the interactive classroom backend through the user side.
  • the SDK can provide a background template preview interface to the app on the user side, so that students can preview the customized templates that can be called through the app; students can also use the background segmentation stickers on the app on the user side to compare The live broadcast background is rendered.
  • the student can also be manually triggered to close.
  • the APP on the user side can report the data of students using stickers to the corresponding back-end (education SaaS back-end or interactive classroom back-end), and the corresponding back-end can analyze which background stickers are used by students and information such as usage amount as additional learning status information.
  • the learning system proposed in the application examples of the present disclosure can not only be applied to online classrooms, but also be extended to other related fields, such as online meetings.
  • the writing order of the steps does not mean a strict execution order but constitutes any limitation on the implementation process.
  • the specific execution order of each step should be based on its function and possibility.
  • the inner logic is determined.
  • the embodiments of the present disclosure also provide a computer-readable storage medium on which computer program instructions are stored, and the computer program instructions implement the above-mentioned method when executed by a processor.
  • the computer-readable storage medium may be a volatile computer-readable storage medium or a non-volatile computer-readable storage medium.
  • An embodiment of the present disclosure also provides an electronic device, including: a processor; a memory for storing executable instructions of the processor; wherein the processor is configured as the above-mentioned method.
  • the embodiment of the present disclosure also provides a computer program, including computer readable code, when the computer readable code is executed in an electronic device, the processor in the electronic device is executed to implement the above method.
  • the above-mentioned memory may be a volatile memory (volatile memory), such as RAM; or a non-volatile memory (non-volatile memory), such as ROM, flash memory, hard disk drive (Hard Disk Drive) , HDD) or solid-state drive (Solid-State Drive, SSD); or a combination of the above types of memory, and provide instructions and data to the processor.
  • volatile memory such as RAM
  • non-volatile memory such as ROM, flash memory, hard disk drive (Hard Disk Drive) , HDD) or solid-state drive (Solid-State Drive, SSD); or a combination of the above types of memory, and provide instructions and data to the processor.
  • the foregoing processor may be at least one of ASIC, DSP, DSPD, PLD, FPGA, CPU, controller, microcontroller, and microprocessor. It is understandable that, for different devices, the electronic device used to implement the above-mentioned processor function may also be other, and the embodiment of the present disclosure does not specifically limit it.
  • the electronic device can be provided as a terminal, server or other form of device.
  • the embodiment of the present disclosure also provides a computer program, which implements the foregoing method when the computer program is executed by a processor.
  • FIG. 4 is a block diagram of an electronic device 800 according to an embodiment of the present disclosure.
  • the electronic device 800 may be a mobile phone, a computer, a digital broadcasting terminal, a messaging device, a game console, a tablet device, a medical device, a fitness device, a personal digital assistant, and other terminals.
  • the electronic device 800 may include one or more of the following components: a processing component 802, a memory 804, a power component 806, a multimedia component 808, an audio component 810, an input/output (I/O) interface 812, and a sensor component 814 , And communication component 816.
  • the processing component 802 generally controls the overall operations of the electronic device 800, such as operations associated with display, telephone calls, data communications, camera operations, and recording operations.
  • the processing component 802 may include one or more processors 820 to execute instructions to complete all or part of the steps of the foregoing method.
  • the processing component 802 may include one or more modules to facilitate the interaction between the processing component 802 and other components.
  • the processing component 802 may include a multimedia module to facilitate the interaction between the multimedia component 808 and the processing component 802.
  • the memory 804 is configured to store various types of data to support operations in the electronic device 800. Examples of these data include instructions for any application or method to operate on the electronic device 800, contact data, phone book data, messages, pictures, videos, etc.
  • the memory 804 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable and Programmable Read Only Memory (EPROM), Programmable Read Only Memory (PROM), Read Only Memory (ROM), Magnetic Memory, Flash Memory, Magnetic Disk or Optical Disk.
  • SRAM static random access memory
  • EEPROM electrically erasable programmable read-only memory
  • EPROM erasable and Programmable Read Only Memory
  • PROM Programmable Read Only Memory
  • ROM Read Only Memory
  • Magnetic Memory Flash Memory
  • Magnetic Disk Magnetic Disk or Optical Disk.
  • the power supply component 806 provides power for various components of the electronic device 800.
  • the power supply component 806 may include a power management system, one or more power supplies, and other components associated with the generation, management, and distribution of power for the electronic device 800.
  • the multimedia component 808 includes a screen that provides an output interface between the electronic device 800 and the user.
  • the screen may include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen may be implemented as a touch screen to receive input signals from the user.
  • the touch panel includes one or more touch sensors to sense touch, sliding, and gestures on the touch panel. The touch sensor may not only sense the boundary of a touch or slide action, but also detect the duration and pressure related to the touch or slide operation.
  • the multimedia component 808 includes a front camera and/or a rear camera. When the electronic device 800 is in an operation mode, such as a shooting mode or a video mode, the front camera and/or the rear camera can receive external multimedia data. Each front camera and rear camera can be a fixed optical lens system or have focal length and optical zoom capabilities.
  • the audio component 810 is configured to output and/or input audio signals.
  • the audio component 810 includes a microphone (MIC), and when the electronic device 800 is in an operation mode, such as a call mode, a recording mode, and a voice recognition mode, the microphone is configured to receive an external audio signal.
  • the received audio signal may be further stored in the memory 804 or transmitted via the communication component 816.
  • the audio component 810 further includes a speaker for outputting audio signals.
  • the I/O interface 812 provides an interface between the processing component 802 and a peripheral interface module.
  • the above-mentioned peripheral interface module may be a keyboard, a click wheel, a button, and the like. These buttons may include, but are not limited to: home button, volume button, start button, and lock button.
  • the sensor component 814 includes one or more sensors for providing the electronic device 800 with various aspects of state evaluation.
  • the sensor component 814 can detect the on/off status of the electronic device 800 and the relative positioning of the components.
  • the component is the display and the keypad of the electronic device 800.
  • the sensor component 814 can also detect the electronic device 800 or the electronic device 800.
  • the position of the component changes, the presence or absence of contact between the user and the electronic device 800, the orientation or acceleration/deceleration of the electronic device 800, and the temperature change of the electronic device 800.
  • the sensor component 814 may include a proximity sensor configured to detect the presence of nearby objects when there is no physical contact.
  • the sensor component 814 may also include a light sensor, such as a CMOS or CCD image sensor, for use in imaging applications.
  • the sensor component 814 may also include an acceleration sensor, a gyroscope sensor, a magnetic sensor, a pressure sensor, or a temperature sensor.
  • the communication component 816 is configured to facilitate wired or wireless communication between the electronic device 800 and other devices.
  • the electronic device 800 can access a wireless network based on a communication standard, such as WiFi, 2G, 3G, 4G, or 5G, or a combination thereof.
  • the communication component 816 receives a broadcast signal or broadcast related personnel information from an external broadcast management system via a broadcast channel.
  • the communication component 816 further includes a near field communication (NFC) module to facilitate short-range communication.
  • the NFC module can be implemented based on radio frequency identification (RFID) technology, infrared data association (IrDA) technology, ultra-wideband (UWB) technology, Bluetooth (BT) technology and other technologies.
  • RFID radio frequency identification
  • IrDA infrared data association
  • UWB ultra-wideband
  • Bluetooth Bluetooth
  • the electronic device 800 may be implemented by one or more application-specific integrated circuits (ASIC), digital signal processors (DSP), digital signal processing devices (DSPD), programmable logic devices (PLD), field-available A programmable gate array (FPGA), controller, microcontroller, microprocessor, or other electronic components are implemented to implement the above methods.
  • ASIC application-specific integrated circuits
  • DSP digital signal processors
  • DSPD digital signal processing devices
  • PLD programmable logic devices
  • FPGA field-available A programmable gate array
  • controller microcontroller, microprocessor, or other electronic components are implemented to implement the above methods.
  • a non-volatile computer-readable storage medium such as the memory 804 including computer program instructions, which can be executed by the processor 820 of the electronic device 800 to complete the foregoing method.
  • FIG. 5 is a block diagram of an electronic device 1900 according to an embodiment of the present disclosure.
  • the electronic device 1900 may be provided as a server. 5
  • the electronic device 1900 includes a processing component 1922, which further includes one or more processors, and a memory resource represented by a memory 1932, for storing instructions executable by the processing component 1922, such as application programs.
  • the application program stored in the memory 1932 may include one or more modules each corresponding to a set of instructions.
  • the processing component 1922 is configured to execute instructions to perform the above-described methods.
  • the electronic device 1900 may also include a power supply component 1926 configured to perform power management of the electronic device 1900, a wired or wireless network interface 1950 configured to connect the electronic device 1900 to the network, and an input output (I/O) interface 1958 .
  • the electronic device 1900 can operate based on an operating system stored in the memory 1932, such as Windows ServerTM, Mac OS XTM, UnixTM, LinuxTM, FreeBSDTM or the like.
  • a non-volatile computer-readable storage medium is also provided, such as the memory 1932 including computer program instructions, which can be executed by the processing component 1922 of the electronic device 1900 to complete the foregoing method.
  • the present disclosure may be a system, method and/or computer program product.
  • the computer program product may include a computer-readable storage medium loaded with computer-readable program instructions for enabling a processor to implement various aspects of the present disclosure.
  • the computer-readable storage medium may be a tangible device that can hold and store instructions used by the instruction execution device.
  • the computer-readable storage medium may be, for example, but not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing.
  • Non-exhaustive list of computer-readable storage media include: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM) Or flash memory), static random access memory (SRAM), portable compact disk read-only memory (CD-ROM), digital versatile disk (DVD), memory stick, floppy disk, mechanical encoding device, such as a printer with instructions stored thereon
  • RAM random access memory
  • ROM read-only memory
  • EPROM erasable programmable read-only memory
  • flash memory flash memory
  • SRAM static random access memory
  • CD-ROM compact disk read-only memory
  • DVD digital versatile disk
  • memory stick floppy disk
  • mechanical encoding device such as a printer with instructions stored thereon
  • the computer-readable storage medium used here is not interpreted as the instantaneous signal itself, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (for example, light pulses through fiber optic cables), or through wires Transmission of electrical signals.
  • the computer-readable program instructions described herein can be downloaded from a computer-readable storage medium to various computing/processing devices, or downloaded to an external computer or external storage device via a network, such as the Internet, a local area network, a wide area network, and/or a wireless network.
  • the network may include copper transmission cables, optical fiber transmission, wireless transmission, routers, firewalls, switches, gateway computers, and/or edge servers.
  • the network adapter card or network interface in each computing/processing device receives computer-readable program instructions from the network, and forwards the computer-readable program instructions for storage in the computer-readable storage medium in each computing/processing device .
  • the computer program instructions used to perform the operations of the present disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-related instructions, microcode, firmware instructions, state setting data, or in one or more programming languages.
  • Source code or object code written in any combination, the programming language includes object-oriented programming languages such as Smalltalk, C++, etc., and conventional procedural programming languages such as "C" language or similar programming languages.
  • Computer-readable program instructions can be executed entirely on the user's computer, partly on the user's computer, executed as a stand-alone software package, partly on the user's computer and partly executed on a remote computer, or entirely on the remote computer or server implement.
  • the remote computer can be connected to the user's computer through any kind of network-including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computer (for example, using an Internet service provider to connect to the user's computer) connect).
  • LAN local area network
  • WAN wide area network
  • an electronic circuit such as a programmable logic circuit, a field programmable gate array (FPGA), or a programmable logic array (PLA), is personalized by using status personnel information of computer-readable program instructions.
  • FPGA field programmable gate array
  • PDA programmable logic array
  • the computer-readable program instructions can be executed to implement various aspects of the present disclosure.
  • These computer-readable program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, thereby producing a machine that makes these instructions when executed by the processor of the computer or other programmable data processing device , A device that implements the functions/actions specified in one or more blocks in the flowcharts and/or block diagrams is produced. It is also possible to store these computer-readable program instructions in a computer-readable storage medium. These instructions make computers, programmable data processing apparatuses, and/or other devices work in a specific manner. Thus, the computer-readable medium storing the instructions includes An article of manufacture, which includes instructions for implementing various aspects of the functions/actions specified in one or more blocks in the flowcharts and/or block diagrams.
  • each block in the flowchart or block diagram may represent a module, program segment, or part of an instruction, and the module, program segment, or part of an instruction contains one or more components for realizing the specified logical function.
  • Executable instructions may also occur in a different order from the order marked in the drawings. For example, two consecutive blocks can actually be executed substantially in parallel, or they can sometimes be executed in the reverse order, depending on the functions involved.
  • each block in the block diagram and/or flowchart, and the combination of the blocks in the block diagram and/or flowchart can be implemented by a dedicated hardware-based system that performs the specified functions or actions Or it can be realized by a combination of dedicated hardware and computer instructions.

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Multimedia (AREA)
  • Strategic Management (AREA)
  • Tourism & Hospitality (AREA)
  • Educational Technology (AREA)
  • Educational Administration (AREA)
  • General Health & Medical Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Primary Health Care (AREA)
  • Economics (AREA)
  • Human Resources & Organizations (AREA)
  • Marketing (AREA)
  • General Business, Economics & Management (AREA)
  • Human Computer Interaction (AREA)
  • Software Systems (AREA)
  • Computational Linguistics (AREA)
  • Social Psychology (AREA)
  • Psychiatry (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Image Analysis (AREA)
  • Electrically Operated Instructional Devices (AREA)

Abstract

The present disclosure relates to a video processing method and apparatus, and an electronic device and a storage medium. The method comprises: acquiring a video, wherein at least some video frames in the video include a target object; according to the video, detecting at least one type of learning behavior conducted by the target object in the process of watching a teaching course; and upon detecting that the target object conducts at least one type of learning behavior, generating learning state information according to at least some video frames which include the at least one type of learning behavior and/or the length of time in which the target object conducts the at least one type of learning behavior.

Description

视频处理方法及装置、电子设备和存储介质Video processing method and device, electronic equipment and storage medium
本公开要求在2020年05月22日提交中国专利局、申请号为202010442733.6、申请名称为“视频处理方法及装置、电子设备和存储介质”的中国专利申请的优先权，其全部内容通过引用结合在本公开中。This disclosure requires the priority of a Chinese patent application filed with the Chinese Patent Office, the application number is 202010442733.6, and the application name is "video processing method and device, electronic equipment and storage medium" on May 22, 2020, the entire content of which is incorporated by reference In this disclosure.
技术领域Technical field
本公开涉及计算机视觉领域，尤其涉及一种视频处理方法及装置、电子设备和存储介质。The present disclosure relates to the field of computer vision, and in particular to a video processing method and device, electronic equipment, and storage medium.
背景技术Background technique
在教学过程中，由于老师需要集中精力授课，使得机构或者老师难以掌握学生的听课状态，家长也无法了解孩子在学校的表现。学生是否真正上课以及是否在认真听课、课堂互动表现如何，都无法量化评估。In the teaching process, because the teacher needs to concentrate on teaching, it is difficult for the institution or the teacher to grasp the student's listening status, and the parents cannot understand the child's performance in school. Whether students are actually attending the class, whether they are listening to the class carefully, and how well they perform in class interactions cannot be quantitatively evaluated.
因此，如何在保证教学质量的同时，掌握每个学生在教学过程中的学习状态，成为目前一个亟待解决的问题。Therefore, how to grasp the learning status of each student in the teaching process while ensuring the quality of teaching has become an urgent problem to be solved at present.
发明内容Summary of the invention
本公开提出了一种视频处理的方案。The present disclosure proposes a video processing solution.
根据本公开的一方面，提供了一种视频处理方法，包括：According to an aspect of the present disclosure, there is provided a video processing method, including:
获取视频，其中，所述视频中的至少部分视频帧包含目标对象；根据所述视频，对所述目标对象在观看教学课程过程中的至少一类学习行为进行检测；在检测到所述目标对象执行至少一类学习行为的情况下，根据至少部分包含所述至少一类学习行为的视频帧和/或所述目标对象执行所述至少一类学习行为的持续时间，生成学习状态信息。Obtain a video, wherein at least part of the video frames in the video contain the target object; according to the video, detect at least one type of learning behavior of the target object during the course of watching the teaching course; when the target object is detected In the case of performing at least one type of learning behavior, the learning state information is generated according to at least a part of the video frame containing the at least one type of learning behavior and/or the duration of the target object performing the at least one type of learning behavior.
根据本公开的一方面，提供了一种视频处理装置，包括：According to an aspect of the present disclosure, there is provided a video processing device, including:
视频获取模块，用于获取视频，其中，所述视频中的至少部分视频帧包含目标对象；A video acquisition module, configured to acquire a video, wherein at least part of the video frames in the video contain the target object;
检测模块，用于根据所述视频，对所述目标对象在观看教学课程过程中的至少一类学习行为进行检测；The detection module is configured to detect at least one type of learning behavior of the target object in the process of watching the teaching course according to the video;
生成模块，用于在检测到所述目标对象执行至少一类学习行为的情况下，根据至少部分包含所述至少一类学习行为的视频帧和/或所述目标对象执行所述至少一类学习行为的持续时间，生成学习状态信息。A generating module, configured to perform the at least one type of learning according to at least part of the video frame containing the at least one type of learning behavior and/or the target object in the case of detecting that the target object performs at least one type of learning behavior The duration of the behavior to generate learning status information.
根据本公开的一方面，提供了一种电子设备，包括：According to an aspect of the present disclosure, there is provided an electronic device including:
处理器；用于存储处理器可执行指令的存储器；其中，所述处理器被配置为：执行上述视频处理方法。A processor; a memory for storing executable instructions of the processor; wherein the processor is configured to execute the above-mentioned video processing method.
根据本公开的一方面，提供了一种计算机可读存储介质，其上存储有计算机程序指令，所述计算机程序指令被处理器执行时实现上述视频处理方法。According to an aspect of the present disclosure, there is provided a computer-readable storage medium having computer program instructions stored thereon, and when the computer program instructions are executed by a processor, the foregoing video processing method is implemented.
根据本公开的一方面，提供了一种计算机程序，包括计算机可读代码，当所述计算机可读代码在电子设备中运行时，所述电子设备中的处理器执行用于实现上述视频处理方法。According to an aspect of the present disclosure, there is provided a computer program including computer readable code, and when the computer readable code is executed in an electronic device, a processor in the electronic device executes the method for implementing the video processing method described above. .
在本公开实施例中，可以在检测到目标对象存在至少一类学习行为的情况下，利用包含学习行为的视频帧来生成直观的学习状态信息，以及根据学习行为的持续时间来生成量化的学习状态信息，采用上述方式可以灵活地得到具有评估价值的学习状态信息，便于老师或家长等相关人员与机构，有效且准确地掌握学生的学习状态。In the embodiments of the present disclosure, when it is detected that the target object has at least one type of learning behavior, the video frame containing the learning behavior can be used to generate intuitive learning state information, and the quantified learning can be generated according to the duration of the learning behavior Status information. The above-mentioned methods can be used to flexibly obtain learning status information with evaluation value, which is convenient for teachers or parents and other relevant personnel and institutions to effectively and accurately grasp the learning status of students.
应当理解的是，以上的一般描述和后文的细节描述仅是示例性和解释性的，而非限制本公开。根据下面参考附图对示例性实施例的详细说明，本公开的其它特征及方面将变得清楚。It should be understood that the above general description and the following detailed description are only exemplary and explanatory, rather than limiting the present disclosure. According to the following detailed description of exemplary embodiments with reference to the accompanying drawings, other features and aspects of the present disclosure will become clear.
附图说明Description of the drawings
此处的附图被并入说明书中并构成本说明书的一部分，这些附图示出了符合本公开的实施例，并与说明书一起用于说明本公开的技术方案。The drawings herein are incorporated into the specification and constitute a part of the specification. These drawings illustrate embodiments that conform to the present disclosure, and are used together with the specification to explain the technical solutions of the present disclosure.
图1示出根据本公开一实施例的视频处理方法的流程图。Fig. 1 shows a flowchart of a video processing method according to an embodiment of the present disclosure.
图2示出根据本公开一实施例的视频处理装置的框图。Fig. 2 shows a block diagram of a video processing device according to an embodiment of the present disclosure.
图3示出根据本公开一应用示例的示意图。Fig. 3 shows a schematic diagram of an application example according to the present disclosure.
图4示出根据本公开实施例的一种电子设备的框图。Fig. 4 shows a block diagram of an electronic device according to an embodiment of the present disclosure.
图5示出根据本公开实施例的一种电子设备的框图。Fig. 5 shows a block diagram of an electronic device according to an embodiment of the present disclosure.
具体实施方式Detailed ways
以下将参考附图详细说明本公开的各种示例性实施例、特征和方面。附图中相同的附图标记表示功能相同或相似的元件。尽管在附图中示出了实施例的各种方面，但是除非特别指出，不必按比例绘制附图。Various exemplary embodiments, features, and aspects of the present disclosure will be described in detail below with reference to the drawings. The same reference numerals in the drawings indicate elements with the same or similar functions. Although various aspects of the embodiments are shown in the drawings, unless otherwise noted, the drawings are not necessarily drawn to scale.
在这里专用的词“示例性”意为“用作例子、实施例或说明性”。这里作为“示例性”所说明的任何实施例不必解释为优于或好于其它实施例。The dedicated word "exemplary" here means "serving as an example, embodiment, or illustration." Any embodiment described herein as "exemplary" need not be construed as being superior or better than other embodiments.
本文中术语“和/或”，仅仅是一种描述关联对象的关联关系，表示可以存在三种关系，例如，A和/或B，可以表示：单独存在A，同时存在A和B，单独存在B这三种情况。另外，本文中术语“至少一种”表示多种中的任意一种或多种中的至少两种的任意组合，例如，包括A、B、C中的至少一种，可以表示包括从A、B和C构成的集合中选择的任意一个或多个元素。The term "and/or" in this article is only an association relationship describing the associated objects, which means that there can be three relationships, for example, A and/or B, which can mean: A alone exists, A and B exist at the same time, exist alone B these three situations. In addition, the term "at least one" in this document means any one of a plurality of or any combination of at least two of the plurality, for example, including at least one of A, B, and C, may mean including A, Any one or more elements selected in the set formed by B and C.
另外，为了更好地说明本公开，在下文的具体实施方式中给出了众多的具体细节。本领域技术人员应当理解，没有某些具体细节，本公开同样可以实施。在一些实例中，对于本领域技术人员熟知的方法、手段、元件和电路未作详细描述，以便于凸显本公开的主旨。In addition, in order to better illustrate the present disclosure, numerous specific details are given in the following specific embodiments. Those skilled in the art should understand that the present disclosure can also be implemented without certain specific details. In some instances, the methods, means, elements, and circuits well known to those skilled in the art have not been described in detail, so as to highlight the gist of the present disclosure.
图1示出根据本公开一实施例的视频处理方法的流程图，该方法可以应用于视频处理装置，视频处理装置可以为终端设备、服务器或者其他处理设备等。其中，终端设备可以为用户设备(User Equipment，UE)、移动设备、用户终端、终端、蜂窝电话、无绳电话、个人数字助理(Personal Digital Assistant，PDA)、手持设备、计算设备、车载设备、可穿戴设备等。在一个示例中，该数据处理方法可以应用于云端服务器或本地服务器，云端服务器可以为公有云服务器，也可以为私有云服务器，根据实际情况灵活选择即可。Fig. 1 shows a flowchart of a video processing method according to an embodiment of the present disclosure. The method can be applied to a video processing device, and the video processing device can be a terminal device, a server, or other processing equipment. Among them, terminal devices can be User Equipment (UE), mobile devices, user terminals, terminals, cellular phones, cordless phones, personal digital assistants (Personal Digital Assistants, PDAs), handheld devices, computing devices, vehicle-mounted devices, and mobile devices. Wearable equipment, etc. In an example, the data processing method can be applied to a cloud server or a local server, and the cloud server can be a public cloud server or a private cloud server, which can be flexibly selected according to actual conditions.
在一些可能的实现方式中，该视频处理方法也可以通过处理器调用存储器中存储的计算机可读指令的方式来实现。In some possible implementations, the video processing method can also be implemented by a processor invoking computer-readable instructions stored in the memory.
如图1所示，在一种可能的实现方式中，所述视频处理方法可以包括：As shown in FIG. 1, in a possible implementation manner, the video processing method may include:
步骤S11，获取视频，其中，视频中的至少部分视频帧包含目标对象。Step S11: Obtain a video, where at least part of the video frames in the video contain the target object.
步骤S12，根据视频，对目标对象在观看教学课程过程中的至少一类学习行为进行检测。Step S12, according to the video, detect at least one type of learning behavior of the target object in the process of watching the teaching course.
步骤S13，在检测到目标对象执行至少一类学习行为的情况下，根据至少部分包含至少一类学习行为的视频帧和/或目标对象执行至少一类学习行为的持续时间，生成学习状态信息。Step S13, in a case where it is detected that the target object performs at least one type of learning behavior, generate learning state information according to at least part of the video frames containing at least one type of learning behavior and/or the duration of the target object performing at least one type of learning behavior.
其中，目标对象可以是任意被获取学习状态信息的对象，即具有学习状态评价需求的对象，其具体实现形式可以根据实际情况灵活确定。在一种可能的实现方式中，目标对象可以是学生，比如小学生、中学生或大学生等；在一种可能的实现方式中，目标对象可以是进修的成年人，比如参加职业教育培训的成年人，或是在老年大学中学习的老年人等。Among them, the target object can be any object whose learning status information is acquired, that is, an object with learning status evaluation requirements, and its specific implementation form can be flexibly determined according to actual conditions. In a possible realization method, the target object may be students, such as elementary school students, middle school students or college students, etc.; in a possible realization method, the target object may be adults with advanced studies, such as adults participating in vocational education and training. Or the elderly who study in senior colleges, etc.
本公开实施例中，视频可以是目标对象在观看教学课程过程中所录制的视频，其中，教学课程的实现形式不受限定，可以是预先录制的课程视频，也可以是直播课程或是教师现场授课的课程等；视频中的至少部分视频帧可以包含目标对象，即录制的视频中目标对象的出现情况可以根据实际情况灵活决定。在一种可能的实现方式中，目标对象可以一直在视频中，在一种可能的实现方式中，目标对象也可以在某些时刻或某些时间段，未出现在视频帧等。In the embodiments of the present disclosure, the video may be a video recorded by the target object while watching the teaching course. The realization form of the teaching course is not limited. It may be a pre-recorded course video, or a live course or a teacher’s site. Courses taught, etc.; at least part of the video frames in the video can contain the target object, that is, the appearance of the target object in the recorded video can be flexibly determined according to the actual situation. In a possible implementation manner, the target object may always be in the video. In a possible implementation manner, the target object may also not appear in the video frame at certain moments or certain periods of time.
目标对象观看教学课程的场景可以根据实际情况灵活决定，在一种可能的实现方式中，这一场景可以为线上场景，即目标对象通过网络课堂等在线教育方式观看教学课程等；在一种可能的实现方式中，这一场景也可以为线下场景，即目标对象通过传统的面对面授课的方式来观看老师当场传授的教学课程，或是目标对象在教室等特定的教学场所中观看通过视频或其他多媒体形式所播放的教学课程等。The scene in which the target object views the teaching course can be flexibly determined according to the actual situation. In a possible implementation, this scene can be an online scene, that is, the target object watches the teaching course through online education methods such as online classrooms; Among possible implementations, this scene can also be an offline scene, that is, the target object can watch the teaching course taught by the teacher on the spot through the traditional face-to-face teaching method, or the target object can watch the video through the classroom and other specific teaching places. Or teaching courses played in other multimedia formats.
视频的具体实现形式可以根据视频处理方法的应用场景灵活决定。在一种可能的实现方式中，视频可以是实时视频，比如目标对象通过在线课堂学习的过程中所实时录制的视频，或是目标对象在教室上课的过程中，通过部署在教室的摄像头来采集到的实时视频等；在一种可能的实现方式中，视 频也可以是录制视频，比如目标对象通过在线课堂学习后，录制到的目标对象学习的回放视频，或是目标对象在教室上课后，通过部署在教室的摄像头来采集到的完整课堂学习视频等。The specific implementation form of the video can be flexibly determined according to the application scenario of the video processing method. In one possible implementation, the video can be a real-time video, such as a video recorded in real time by the target object in the online classroom learning process, or the target object is captured by a camera deployed in the classroom while the target object is in the classroom. In one possible implementation, the video can also be a recorded video. For example, after the target object learns through an online classroom, the recorded playback video of the target object learning, or the target object after the classroom has a class, Complete classroom learning videos collected through cameras deployed in the classroom.
为了便于描述，后续各公开实施例均以视频是目标对象通过在线课堂学习的过程中所实时录制的视频为例，对视频处理过程进行说明。其他应用场景下的视频处理过程可以参考后续各公开实施例进行灵活扩展，在此不再赘述。For ease of description, the subsequent disclosed embodiments all take the video recorded in real time during the online classroom learning process of the target object as an example to illustrate the video processing process. The video processing process in other application scenarios can be flexibly extended with reference to the subsequent disclosed embodiments, which will not be repeated here.
在通过步骤S11获取如上述各公开实施例所述的视频后，可以通过步骤S12，对目标对象在观看教学课程过程中的至少一类学习行为进行检测。其中被检测的学习行为的种类和数量可以根据实际情况灵活决定，不局限于下述各公开实施例。在一种可能的实现方式中，目标对象执行的学习行为可以包括以下行为中的至少一类：比如执行至少一种目标手势、表现目标情绪、关注教学课程的展示区域、与其他对象产生至少一种互动行为、在视频中的至少部分视频帧中未出现、闭眼以及在教学课程的展示区域内的目光交流等。After obtaining the video as described in the above-mentioned disclosed embodiments in step S11, step S12 can be used to detect at least one type of learning behavior of the target object in the process of watching the teaching course. The type and quantity of the detected learning behaviors can be flexibly determined according to actual conditions, and are not limited to the following disclosed embodiments. In a possible implementation manner, the learning behavior performed by the target object may include at least one of the following behaviors: for example, performing at least one target gesture, expressing the target emotion, paying attention to the display area of the teaching course, and generating at least one behavior with other objects. This kind of interactive behavior, not appearing in at least part of the video frame in the video, closing eyes, and eye contact in the display area of the teaching course.
其中，目标手势可以是反映目标对象在观看教学课程过程中，可能会产生的某些预设手势，其具体实现形式可以根据实际情况灵活设定，详见后续各公开实施例，在此先不做展开。Among them, the target gesture can reflect certain preset gestures that the target object may produce during the course of watching the teaching course. The specific implementation form can be flexibly set according to the actual situation. For details, please refer to the subsequent disclosed embodiments. Do unfold.
目标情绪可以是目标对象在观看教学课程过程中，反映出对教学课程真实感受的某些情绪，其具体实现形式同样可以根据实际情况灵活设定，在此先不做展开。The target emotion can be some emotions of the target object that reflect the true feelings of the teaching course during the process of watching the teaching course. The specific realization form can also be flexibly set according to the actual situation, and will not be expanded here.
关注教学课程的展示区域可以体现目标对象在观看教学课程的过程中的关注度，其中，展示区域的具体区域范围可以根据实际情况灵活设定，不局限于下述各公开实施例。在一种可能的实现方式中，展示区域可以是线上课堂中教学课程视频的展示区域，比如学生通过电脑、手机或是平板电脑等终端设备进行在线学习的过程中，展示区域可以是这些终端设备中播放教学课程的屏幕等；在一种可能的实现方式中，展示区域可以是线下课堂中教师的教学区域，比如教室中的讲台或是黑板等位置。Focusing on the display area of the teaching course can reflect the attention of the target object in the process of watching the teaching course. The specific area range of the display area can be flexibly set according to the actual situation and is not limited to the following disclosed embodiments. In a possible implementation, the display area can be the display area of the teaching course video in the online classroom. For example, when students are learning online through terminal devices such as computers, mobile phones, or tablets, the display area can be these terminals The screen for playing the teaching course in the device, etc.; in a possible implementation, the display area may be the teaching area of the teacher in the offline classroom, such as the podium or blackboard in the classroom.
与其他对象产生至少一种互动行为可以是目标对象在观看教学课程过程中，与教学课程中相关的其他对象所产生的与学习相关的互动，其中，其他对象的实现形式可以根据实际情况灵活决定，在一种可能的实现方式中，其他对象可以是授课对象，比如教师等，在一种可能的实现方式中，其他对象也可以是教学过程中除目标对象以外的学习对象，比如目标对象的同学等；与其他对象产生的互动行为可以根据对象的不同灵活发生变化，在一种可能的实现方式中，在其他对象为授课教师的情况下，与其他对象产生的互动可以包括接收老师所发送的奖励，比如收到老师发的小红花或点名表扬等，在一种可能的实现方式中，在其他对象为授课教师的情况下，与其他对象产生的互动可以包括回答老师的问题或根据老师的点名进行发言等，在一种可能的实现方式中，在其他对象为同学的情况下，与其他对象产生的互动可以包括小组互助、小组讨论或小组学习等。At least one kind of interaction behavior with other objects can be the learning-related interaction generated by the target object and other objects related to the teaching course during the course of watching the teaching course. Among them, the realization form of other objects can be flexibly determined according to the actual situation. In a possible implementation, other objects can be teaching objects, such as teachers, etc. In a possible implementation, other objects can also be learning objects other than the target object in the teaching process, such as the target object’s Classmates, etc.; the interaction behavior with other objects can be flexibly changed according to the different objects. In a possible implementation mode, when the other objects are the instructors, the interaction with other objects can include receiving sent by the teacher For example, receiving small red flowers from the teacher or commendation by name, etc., in a possible implementation mode, when other objects are the instructors, the interaction with other objects can include answering the teacher’s questions or according to In a possible implementation manner, when the other objects are students, the interaction with other objects can include group mutual assistance, group discussion, or group study.
在视频中的至少部分视频帧中未出现可以是学习对象在某些时刻或某些时间段，离开了教学课程等情况，比如目标对象在线学习的过程中，可能由于私人原因暂时离开当前的在线学习设备，或是离开当前的在线学习设备的拍摄范围内等。At least part of the video frame does not appear in the video. It may be that the learning object has left the teaching course at certain moments or certain time periods. For example, the target object may temporarily leave the current online due to personal reasons during the online learning process. Learning equipment, or leaving the shooting range of the current online learning equipment, etc.
闭眼可以是目标对象在观看教学课程的过程中进行的闭眼操作，在教学课程的展示区域内的目光交流，可以是观看教学课程的展示区域，与之相对应的，根据视频中目标对象在教学课程的展示区域内的目光交流的情况，还可以进一步确定目标对象未观看教学课程的展示区域的情况等。Eye-closing can be the closed-eye operation performed by the target object in the process of watching the teaching course. The eye contact in the display area of the teaching course can be the display area of watching the teaching course. Correspondingly, according to the target object in the video The situation of eye contact in the display area of the teaching course can further determine the situation that the target object has not watched the display area of the teaching course, etc.
通过上述公开实施例中提到的各类学习行为，可以对目标对象的学习过程进行全面且灵活的行为检测，从而提升根据检测所得到的学习状态信息的全面性和准确性，更加灵活准确地掌握目标对象的学习状态。Through the various learning behaviors mentioned in the above disclosed embodiments, comprehensive and flexible behavior detection can be performed on the learning process of the target object, thereby improving the comprehensiveness and accuracy of the learning status information obtained according to the detection, and making it more flexible and accurate. Grasp the learning status of the target object.
具体地，步骤S12对上述公开实施例中的各类学习行为执行哪类或哪几类检测，可以根据实际情况灵活设定。在一种可能的实现方式中，可以对上述公开实施例中提到的各类学习行为同时进行检测，具体的检测方式与过程可以详见下述各公开实施例，在此先不做展开。Specifically, in step S12, which type or types of detections are performed on the various learning behaviors in the above disclosed embodiment can be flexibly set according to actual conditions. In a possible implementation manner, the various learning behaviors mentioned in the above disclosed embodiments can be detected at the same time, and the specific detection methods and processes can be detailed in the following disclosed embodiments, which will not be expanded here.
在检测到目标对象执行至少一类学习行为的情况下，可以根据至少部分包含至少一类学习行为的视频帧和/或目标对象执行至少一类学习行为的持续时间，来生成学习状态信息。其中，学习状态信息的具体实现形式，可以根据学习行为的种类，以及对应执行的操作所灵活决定。在一种可能的实现方式中，在根据至少部分包含至少一类学习行为的视频帧生成学习状态信息的情况下，学习状态信 息可以包括由视频帧所组成的信息；在一种可能的实现方式中，在根据目标对象执行至少一类学习行为的持续时间的情况下，学习状态信息可以为数字形式的数据信息；在一种可能的实现方式中，学习状态信息也可以同时包含有视频帧信息和数据信息这两种形式的信息；在一种可能的实现方式中，学习状态信息也可以包含其他状态的信息等。具体地，如何生成学习状态信息以及学习状态信息的实现形式，可以参考后续各公开实施例，在此先不做展开。In the case where it is detected that the target object performs at least one type of learning behavior, the learning state information may be generated according to a video frame containing at least part of the at least one type of learning behavior and/or the duration of the target object performing at least one type of learning behavior. Among them, the specific implementation form of the learning status information can be flexibly determined according to the type of learning behavior and the corresponding operation performed. In a possible implementation manner, in a case where the learning state information is generated based on video frames that at least partially contain at least one type of learning behavior, the learning state information may include information composed of video frames; in a possible implementation manner In the case of performing at least one type of learning behavior according to the duration of the target object, the learning state information may be data information in digital form; in a possible implementation, the learning state information may also include video frame information at the same time And data information; in a possible implementation, the learning status information can also include other status information. Specifically, how to generate the learning state information and the implementation form of the learning state information can refer to the subsequent disclosed embodiments, and will not be expanded here.
在本公开实施例中，可以在检测到目标对象存在至少一类学习行为的情况下，利用包含学习行为的视频帧来生成直观的学习状态信息，以及根据学习行为的持续时间来生成量化的学习状态信息，采用上述方式可以灵活地得到具有评估价值的学习状态信息，便于老师或家长等相关人员与机构，有效且准确地掌握学生的学习状态。In the embodiments of the present disclosure, when it is detected that the target object has at least one type of learning behavior, the video frame containing the learning behavior can be used to generate intuitive learning state information, and the quantified learning can be generated according to the duration of the learning behavior Status information. The above-mentioned methods can be used to flexibly obtain learning status information with evaluation value, which is convenient for teachers or parents and other relevant personnel and institutions to effectively and accurately grasp the learning status of students.
上述公开实施例中已经提到，视频可以是目标对象在观看教学课程过程中所录制的视频，而目标对象观看教学课程的场景可以根据实际情况灵活决定，因此，相应地，步骤S11中获取视频的方式也可以根据场景不同而灵活发生变化。在一种可能的实现方式中，在目标对象观看教学课程的场景为线上场景的情况下，即目标对象可以通过在线课堂观看教学课程的情况下，获取视频的方式可以包括：如果视频处理装置与目标对象进行在线学习的设备为同一装置，则可以通过目标对象进行在线学习的设备对目标对象观看教学课程的过程进行视频采集；如果视频处理装置与目标对象进行在线学习的设备为不同装置，则可以通过目标对象进行在线学习的设备对目标对象观看教学课程的过程进行视频采集，并实时和/或非实时地传输到视频处理装置中。在一种可能的实现方式中，在目标对象观看教学课程的场景为线下场景的情况下，即目标对象参加面对面授课或是在特定教学场景中观看教学课程视频等情况下，获取视频的方式可以包括：通过部署在线下的图像采集设备(如普通摄像头、应安全需求部署的拍摄器件等)来采集目标对象的视频。进一步地，如果部署在线下的图像采集设备可以进行视频处理，即可以作为视频处理装置，则步骤S11中的获取视频过程已完成；如果部署在线下的图像采集设备无法进行视频处理，则可以将部署在线下的图像采集设备采集的视频实时和/或非实时地传输到视频处理装置中。As mentioned in the above disclosed embodiment, the video can be a video recorded by the target object during watching the teaching course, and the scene of the target object watching the teaching course can be flexibly determined according to the actual situation. Therefore, correspondingly, the video is obtained in step S11 The method can also be flexibly changed according to different scenarios. In a possible implementation manner, in the case where the scene where the target object views the teaching course is an online scene, that is, when the target object can watch the teaching course through the online classroom, the way to obtain the video may include: If the video processing device The device for online learning with the target object is the same device, and the device for online learning for the target object can be used to collect video on the process of the target object watching the teaching course; if the video processing device and the device for online learning for the target object are different devices, Then, the device for online learning of the target object can collect the video of the process of the target object watching the teaching course, and transmit it to the video processing device in real-time and/or non-real-time. In one possible implementation method, when the target object is watching the teaching course in an offline scene, that is, when the target object participates in a face-to-face teaching or watching a teaching course video in a specific teaching scene, the way to obtain the video It may include: collecting the video of the target object by deploying offline image acquisition equipment (such as ordinary cameras, shooting devices deployed in response to security requirements, etc.). Further, if the image acquisition device deployed offline can perform video processing, that is, it can be used as a video processing device, then the video acquisition process in step S11 has been completed; if the image acquisition device deployed offline cannot perform video processing, the The video collected by the offline image acquisition equipment is transmitted to the video processing device in real time and/or non-real time.
如上述各公开实施例所述，步骤S12中对目标对象进行学习行为检测的方式可以根据实际情况灵活决定。在一种可能的实现方式中，步骤S12可以包括：As described in the above disclosed embodiments, the method of performing learning behavior detection on the target object in step S12 can be flexibly determined according to the actual situation. In a possible implementation manner, step S12 may include:
步骤S121，对视频进行目标对象检测，得到包含目标对象的视频帧。Step S121: Perform target object detection on the video to obtain a video frame containing the target object.
步骤S122，对包含目标对象的视频帧进行至少一类学习行为检测。Step S122: Perform at least one type of learning behavior detection on the video frame containing the target object.
通过上述公开实施例可以看出，在一种可能的实现方式中，可以通过对视频进行目标对象检测，确定视频中包含目标对象的视频帧。在确定了哪些视频帧包含有目标对象以后，可以对包含目标对象的视频帧中的目标对象，进行至少一类学习行为检测。It can be seen from the above disclosed embodiments that, in a possible implementation manner, target object detection can be performed on the video to determine the video frame containing the target object in the video. After determining which video frames contain the target object, at least one type of learning behavior detection can be performed on the target object in the video frame containing the target object.
其中，目标对象检测的方式可以根据实际情况灵活决定，不局限于下述实施例。在一种可能的实现方式中，可以通过人脸检测或是人脸跟踪等方式，对视频中的目标对象进行检测。在一种可能的实现方式中，在通过人脸检测或是人脸跟踪等方式对视频帧进行检测后，可能检测到多个对象，在这种情况下，还可以进一步对检测到的人脸图像进行筛选，从中选定一个或多个对象作为目标对象，具体的筛选方式可以根据实际情况灵活设定，在本公开实施例中不做限定。Wherein, the method of detecting the target object can be flexibly determined according to the actual situation, and is not limited to the following embodiments. In a possible implementation manner, the target object in the video can be detected by means such as face detection or face tracking. In a possible implementation, after the video frame is detected by means of face detection or face tracking, multiple objects may be detected. In this case, the detected face may be further detected. The image is screened, and one or more objects are selected as the target object. The specific screening method can be flexibly set according to the actual situation, which is not limited in the embodiment of the present disclosure.
在一种可能的实现方式中，在得到了包含目标对象的视频帧后，可以通过步骤S122，对包含目标对象的视频帧进行至少一类学习行为检测。步骤S122的实现方式可以根据学习行为的不同而灵活发生变化，详见下述各公开实施例，在此先不做展开。在需要对目标对象的多类学习行为进行检测的情况下，可以同时采用多种方式进行组合来实现多类学习行为检测。In a possible implementation manner, after the video frame containing the target object is obtained, step S122 may be used to perform at least one type of learning behavior detection on the video frame containing the target object. The implementation of step S122 can be flexibly changed according to different learning behaviors. For details, refer to the following disclosed embodiments, which will not be expanded here. In the case where multiple types of learning behaviors of the target object need to be detected, multiple methods can be combined to achieve multiple types of learning behavior detection at the same time.
在一些可能的实现方式中，在对视频进行目标对象检测后，即可以完成对目标对象在观看教学课程过程中的学习行为检测。即可以通过对视频进行目标对象检测，确定上述公开实施例中提到的在视频中的至少部分视频帧中未出现这一学习行为。并进一步根据未检测到目标对象的视频帧，来得到学习状态信息，或是根据未检测到目标对象的视频帧来统计目标对象在视频中的至少部分视频帧中未出现的时间作为学习状态信息。In some possible implementation manners, after the target object is detected on the video, the learning behavior detection of the target object in the process of watching the teaching course can be completed. That is, by performing target object detection on the video, it can be determined that this learning behavior does not appear in at least part of the video frames in the video mentioned in the above-mentioned disclosed embodiment. And further obtain the learning state information according to the video frame of the undetected target object, or calculate the time that the target object does not appear in at least some of the video frames in the video according to the video frame of the undetected target object as the learning state information .
在本公开实施例中，通过对视频进行目标对象检测，得到包含目标对象的视频帧，以及对包含 目标对象的视频帧进行至少一类学习行为检测，通过上述过程，可以利用对视频进行的目标对象检测，更加有针对性地对目标对象的至少一类学习行为进行检测，从而使得学习行为检测更为准确，进一步提高后续得到的学习状态信息的准确性与可靠性。In the embodiments of the present disclosure, by performing target object detection on the video, a video frame containing the target object is obtained, and at least one type of learning behavior detection is performed on the video frame containing the target object. Through the above process, the target object of the video can be used. Object detection is more targeted to detect at least one type of learning behavior of the target object, thereby making the learning behavior detection more accurate, and further improving the accuracy and reliability of the subsequent learning state information.
如上述各公开实施例所述，步骤S122的实现方式可以根据学习行为的不同而灵活发生变化。在一种可能的实现方式中，学习行为可以包括：执行至少一种目标手势；As described in the above disclosed embodiments, the implementation of step S122 can be flexibly changed according to different learning behaviors. In a possible implementation manner, the learning behavior may include: performing at least one target gesture;
在这种情况下，对包含目标对象的视频帧进行至少一类学习行为检测，可以包括：In this case, performing at least one type of learning behavior detection on the video frame containing the target object may include:
对包含目标对象的视频帧进行至少一种目标手势的检测；Detecting at least one target gesture on the video frame containing the target object;
在检测到包含至少一种目标手势的连续视频帧的数量超过第一阈值的情况下，将包含目标手势的视频帧中的至少一帧记录为手势开始帧；In a case where it is detected that the number of consecutive video frames containing at least one target gesture exceeds the first threshold, at least one of the video frames containing the target gesture is recorded as a gesture start frame;
在手势开始帧以后的视频帧中，不包含目标手势的连续视频帧的数量超过第二阈值的情况下，将不包含目标手势的视频帧中的至少一帧记录为手势结束帧；In the video frames after the gesture start frame, when the number of consecutive video frames that do not include the target gesture exceeds the second threshold, record at least one of the video frames that do not include the target gesture as the gesture end frame;
根据手势开始帧与手势结束帧的数量，确定视频中所述目标对象执行至少一种目标手势的次数和/或时间。According to the number of gesture start frames and gesture end frames, determine the number of times and/or time for the target object in the video to perform at least one target gesture.
通过上述公开实施例可以看出，在学习行为包括执行至少一种目标手势的情况下，对目标对象的视频帧进行的学习行为检测可以包括目标手势检测。It can be seen from the above disclosed embodiments that, in a case where the learning behavior includes performing at least one target gesture, the learning behavior detection performed on the video frame of the target object may include target gesture detection.
其中，目标手势具体包含哪些手势，可以根据实际情况进行灵活设定，不局限于下述公开实施例。示例性的，目标手势包括举手手势、点赞手势、OK手势以及胜利手势中的一种或多种。Wherein, which gestures the target gesture specifically includes can be flexibly set according to actual conditions, and is not limited to the following disclosed embodiments. Exemplarily, the target gesture includes one or more of a hand-raising gesture, a thumb-up gesture, an OK gesture, and a victory gesture.
在一种可能的实现方式中，目标手势可以包括在观看教学课程的过程中，目标对象根据听课情况所反映的与学习相关的手势，比如用于回答问题的举手手势、对授课内容或授课教师表示赞赏的点赞手势(竖起大拇指等)、对授课内容表示明白或认同的OK手势以及与授课教师之间进行互动的胜利手势(比如Yeah手势等)中等。In a possible implementation, the target gesture can be included in the process of watching the teaching course. The target object reflects the learning-related gestures according to the listening situation, such as the gesture of raising the hand used to answer the question, the content of the lecture or the teaching The teacher's praise gesture (thumbs up, etc.), the OK gesture that expresses understanding or approval of the teaching content, and the victory gesture for interaction with the instructor (such as Yeah gesture, etc.), etc.
具体地，对包含目标对象的视频帧进行至少一种目标手势的检测的方式可以根据实际情况灵活决定，不局限于下述公开实施例。在一种可能的实现方式中，可以通过手势识别的相关算法来实现目标手势的检测，比如可以识别视频帧中目标对象的手部关键点或者手部检测框对应的图像区域，基于手部关键点或手部检测框对应的图像区域进行手势检测，基于手势检测结果确定目标对象是否在执行目标手势。在一种可能的实现方式中，可以通过具有手势检测功能的神经网络，实现目标手势的检测。具有手势检测功能的神经网络的具体结构与实现方式可以根据实际情况进行灵活设定，在目标手势包含多种手势的情况下，在一种可能的实现方式中，可以将包含目标对象的视频帧输入到可以同时检测到多个手势的神经网络，来实现目标手势的检测；在一种可能的实现方式中，也可以将包含目标对象的视频帧分别输入到多个具有单一手势检测功能的神经网络中，来实现多个目标手势的检测。Specifically, the method for detecting at least one target gesture on the video frame containing the target object can be flexibly determined according to the actual situation, and is not limited to the following disclosed embodiments. In a possible implementation, the detection of target gestures can be achieved through the related algorithms of gesture recognition. For example, the key points of the hand of the target object in the video frame or the image area corresponding to the hand detection frame can be recognized, based on the hand key Gesture detection is performed on the image area corresponding to the point or hand detection frame, and based on the gesture detection result, it is determined whether the target object is performing the target gesture. In a possible implementation manner, the detection of the target gesture can be achieved through a neural network with a gesture detection function. The specific structure and implementation of the neural network with gesture detection function can be flexibly set according to the actual situation. In the case that the target gesture includes multiple gestures, in a possible implementation manner, the video frame containing the target object can be Input to the neural network that can detect multiple gestures at the same time to achieve the detection of the target gesture; in a possible implementation, the video frame containing the target object can also be input to multiple neural networks with a single gesture detection function. In the network, to realize the detection of multiple target gestures.
在通过上述任意公开实施例进行目标手势检测的过程中，如果检测到包含至少一种目标手势的连续视频帧的数量超过第一阈值的情况下，可以从这些包含目标手势的连续视频帧中，选定至少一帧作为手势开始帧。其中，第一阈值的数量可以根据实际情况灵活设定，不同的目标手势对应的第一阈值的数量可以相同，也可以不同，比如可以将举手手势对应的第一阈值设定为6，点赞手势对应的第一阈值设定为7，则在检测到包含举手手势的连续视频帧的数量不小于6的情况下，可以从包含举手手势的视频帧中选定至少一帧作为举手手势的手势开始帧，在检测到点赞手势的连续视频帧的数量不小于7的情况下，可以从包含点赞手势的视频帧中选定至少一帧作为点赞手势的手势开始帧。在一种可能的实现方式中，为了便于目标手势的检测，可以将不同目标手势对应的第一阈值设定为同一数值，在一个示例中，第一阈值的数量可以设置为6。In the process of performing target gesture detection through any of the above disclosed embodiments, if it is detected that the number of continuous video frames containing at least one target gesture exceeds the first threshold, these continuous video frames containing the target gesture can be selected, Select at least one frame as the start frame of the gesture. Among them, the number of first thresholds can be flexibly set according to the actual situation. The number of first thresholds corresponding to different target gestures can be the same or different. For example, the first threshold corresponding to the hand-raising gesture can be set to 6. The first threshold corresponding to the thumbs-up gesture is set to 7. If the number of consecutive video frames containing the hand-raising gesture is detected to be not less than 6, at least one frame can be selected from the video frames containing the hand-raising gesture as the gesture. The gesture start frame of the hand gesture, if the number of consecutive video frames of the like gesture is not less than 7, at least one frame may be selected from the video frames containing the like gesture as the gesture start frame of the like gesture. In a possible implementation manner, in order to facilitate the detection of the target gesture, the first thresholds corresponding to different target gestures may be set to the same value. In an example, the number of the first thresholds may be set to 6.
手势开始帧的选定方式同样可以根据实际情况灵活设定，在一种可能的实现方式中，可以将检测到的包含目标手势的连续视频帧中的第一帧，作为该目标手势的手势开始帧，在一种可能的实现方式中，为了减少手势检测的误差，也可以将检测到的包含目标手势的连续视频帧中的第一帧以后的某一帧，作为该目标手势的手势开始帧。The selection method of the gesture start frame can also be flexibly set according to the actual situation. In a possible implementation, the first frame of the detected continuous video frames containing the target gesture can be used as the gesture start of the target gesture Frame, in a possible implementation, in order to reduce the error of gesture detection, a certain frame after the first frame of the detected continuous video frames containing the target gesture can also be used as the gesture start frame of the target gesture .
在确定了手势开始帧以后，可以从手势开始帧以后的视频帧中确定手势结束帧，即确定手势开始帧中的目标手势的结束时间。具体的确定方式可以根据实际情况灵活选择，不局限于下述公开实施 例。在一种可能的实现方式中，可以在检测到手势开始帧以后的视频帧中，检测到不包含手势开始帧中的目标手势的连续视频帧数量超过第二阈值的情况下，将不包含目标手势的连续视频帧中的至少一帧记录为手势结束帧。其中，第二阈值的数值同样可以根据实际情况灵活设定，不同目标手势对应的第二阈值的数值可以相同也可以不同，具体的设定方式可以参考第一阈值，在此不再赘述。在一个示例中，不同目标手势对应的第二阈值的数值可以相同，比如可以设置为10，即在手势开始帧以后，检测到连续10帧不包含手势开始帧中的目标手势的情况下，可以认为目标对象结束执行目标手势。在这种情况下，可以从不包含目标手势的连续视频帧中选定至少一帧作为手势结束帧，选定的方式同样可以参考手势开始帧，在一个示例中，可以将不包含目标手势的连续视频帧中的最后一帧作为手势结束帧；在一个示例中，也可以将不包含目标手势的连续视频帧中的最后一帧以前的某一帧作为手势结束帧。在一种可能的实现方式中，如果在检测到手势开始帧以后，存在某一帧或某几帧不包含目标对象的视频帧，则也可以将不包含目标对象的某一或某些视频帧作为手势结束帧。After the gesture start frame is determined, the gesture end frame can be determined from the video frames after the gesture start frame, that is, the end time of the target gesture in the gesture start frame can be determined. The specific determination method can be flexibly selected according to the actual situation, and is not limited to the following disclosed embodiments. In a possible implementation manner, in the video frames after the gesture start frame is detected, if the number of consecutive video frames that do not contain the target gesture in the gesture start frame exceeds the second threshold, the target will not be included. At least one of the consecutive video frames of the gesture is recorded as the gesture end frame. The value of the second threshold can also be flexibly set according to actual conditions. The values of the second threshold corresponding to different target gestures can be the same or different. The specific setting method can refer to the first threshold, which will not be repeated here. In an example, the value of the second threshold corresponding to different target gestures can be the same, for example, it can be set to 10. That is, after the gesture start frame, it is detected that 10 consecutive frames do not contain the target gesture in the gesture start frame. It is considered that the target object has finished performing the target gesture. In this case, you can select at least one frame from the continuous video frames that do not contain the target gesture as the gesture end frame. The selection method can also refer to the gesture start frame. In one example, you can select the gesture start frame. The last frame in the continuous video frames is used as the gesture end frame; in an example, a frame before the last frame in the continuous video frames that do not contain the target gesture may also be used as the gesture end frame. In a possible implementation, if there is a certain frame or a few frames of video frames that do not contain the target object after the gesture start frame is detected, one or some video frames that do not contain the target object can also be set End the frame as a gesture.
在确定了手势开始帧与手势结束帧以后，可以根据视频帧中包含的手势开始帧与手势结束帧的数量，来确定目标对象执行某种或某些目标手势的次数，进一步地，还可以确定执行某种或某些目标手势的持续时间等。具体确定哪些与目标手势相关的内容，可以根据步骤S13中学习状态信息的需求来灵活决定，详见后续各公开实施例，在此先不做展开。After the gesture start frame and gesture end frame are determined, the number of gesture start frames and gesture end frames contained in the video frame can be used to determine the number of times the target object performs a certain target gesture or certain target gestures. The duration of the execution of a certain or certain target gesture, etc. The specific determination of the content related to the target gesture can be flexibly determined according to the requirement of learning state information in step S13. For details, please refer to the subsequent disclosed embodiments, which will not be expanded here.
通过对包含目标对象的视频帧进行至少一种目标手势的检测，并根据检测情况确定手势开始帧与手势结束帧，从而进一步确定视频中目标对象执行至少一种目标手势的次数和/或时间，通过上述过程，可以对视频中目标对象根据学习状态所反馈的手势进行全面且准确的检测，从而提高后续得到的学习状态信息的全面性与精度，继而可以准确地掌握目标对象的学习状态。By detecting at least one target gesture on the video frame containing the target object, and determining the gesture start frame and the gesture end frame according to the detection situation, thereby further determining the number and/or time of at least one target gesture performed by the target object in the video, Through the above process, it is possible to comprehensively and accurately detect the gestures fed back by the target object in the video according to the learning state, thereby improving the comprehensiveness and accuracy of the subsequent learning state information, and then accurately grasping the learning state of the target object.
在一种可能的实现方式中，学习行为可以包括：表现目标情绪；In a possible implementation, the learning behavior can include: expressing the target emotion;
在这种情况下，对包含目标对象的视频帧进行至少一类学习行为检测，可以包括：In this case, performing at least one type of learning behavior detection on the video frame containing the target object may include:
对包含目标对象的视频帧进行表情检测和/或微笑值检测；Perform expression detection and/or smile value detection on the video frame containing the target object;
在检测到视频帧中目标对象展示至少一种第一目标表情或微笑值检测的结果超过目标微笑值情况下，将检测到的视频帧作为第一检测帧；In a case where it is detected that the target object in the video frame exhibits at least one first target expression or the result of the smile value detection exceeds the target smile value, use the detected video frame as the first detection frame;
在检测到连续的第一检测帧的数量超过第三阈值的情况下，确定目标对象产生目标情绪。In a case where it is detected that the number of consecutive first detection frames exceeds the third threshold, it is determined that the target object generates the target emotion.
其中，目标情绪可以为根据实际需求设定的任意情绪，比如可以为表明目标对象在专注学习的开心情绪，或是表明目标对象学习状态不佳的厌烦情绪等。下述各公开实施例以目标情绪为开心情绪为例进行说明，目标情绪为其他情绪的情况可以参考后续各公开实施例进行相应扩展。Among them, the target emotion can be any emotion set according to actual needs, for example, it can be a happy emotion that indicates that the target object is focused on learning, or a bored emotion that indicates that the target object is in a poor learning state. The following disclosed embodiments are described by taking the target emotion as happy emotion as an example, and the case where the target emotion is other emotions can be expanded with reference to the subsequent disclosed embodiments.
通过上述公开实施例可以看出，在学习行为包括表现目标情绪的情况下，可以通过表情检测和/或微笑值检测，来实现目标对象的学习行为检测。在一种可能的实现方式中，可以仅通过表情检测或微笑值检测来实现表现目标情绪这一学习行为的检测，在一种可能的实现方式中，可以通过表情检测与微笑值检测，来共同确定目标对象是否表现目标情绪。后续各公开实施例均以通过表情检测与微笑值检测来共同确定目标对象是否表现目标情绪为例进行说明，其余实现方式可以参考后续各公开实施例进行相应扩展，在此不再赘述。It can be seen from the above disclosed embodiments that when the learning behavior includes expressing the target emotion, expression detection and/or smile value detection can be used to achieve the learning behavior detection of the target object. In a possible implementation manner, the learning behavior of expressing the target emotion can be detected only by expression detection or smile value detection. In a possible implementation manner, expression detection and smile value detection can be used together. Determine whether the target object expresses the target emotion. The subsequent disclosed embodiments are described by taking as an example the determination of whether the target object expresses the target emotion through expression detection and smile value detection. The remaining implementation manners can be expanded with reference to the subsequent disclosed embodiments, and will not be repeated here.
其中，表情检测可以包括对目标对象展示的表情进行检测，比如可以检测目标对象展示何种表情，具体的表情划分可以根据实际情况灵活设定，在一种可能的实现方式中，可以将表情划分为高兴、平静以及其他等；而微笑值检测可以包括对目标对象的微笑强度进行检测，比如可以检测目标对象的微笑幅度有多大，微笑值检测的结果可以通过数值来反馈，比如可以将微笑值检测的结果设定为在[0,100]之间，数值越高，表明目标对象的微笑强度或是幅度越高等。具体的表情检测与微笑值检测的方式可以根据实际情况灵活决定，任何能检测到目标对象的表情或是微笑程度的方式，均可以作为相应的检测方式，不局限于下述各公开实施例。在一种可能的实现方式中，可以通过表情识别神经网络来实现目标对象的表情检测，在一种可能的实现方式中，可以通过微笑值检测神经网络，来实现目标对象的微笑值检测。具体地表情识别神经网络与微笑值检测神经网络的结构与实现方式在本公开实施例中不做限定，任何可以通过训练实现表情识别功能的神经网络以及通过训练实现微笑值检测功能的神经网络均可以应用于本公开实施例。在一种可能的实现方式中，也可以通过对视频中目标对象的人 脸关键点以及嘴部关键点进行检测，来分别实现表情检测和微笑值检测。Among them, the expression detection can include the detection of the expressions displayed by the target object, for example, it can detect what kind of expression the target object displays. The specific expression division can be flexibly set according to the actual situation. In a possible implementation manner, the expression can be divided For happiness, calmness, etc.; the smile value detection can include the detection of the smile intensity of the target object, for example, it can detect how big the smile of the target object is, and the result of the smile value detection can be fed back by numerical values. The detection result is set to be between [0,100]. The higher the value, the higher the intensity or amplitude of the target's smile. The specific expression detection and smile value detection methods can be flexibly determined according to the actual situation. Any method that can detect the expression or the degree of smile of the target object can be used as a corresponding detection method, and is not limited to the following disclosed embodiments. In a possible implementation manner, the expression detection of the target object can be realized by the facial expression recognition neural network, and in a possible implementation manner, the smile value detection of the target object can be realized by the smile detection neural network. Specifically, the structure and implementation of the facial expression recognition neural network and the smile value detection neural network are not limited in the embodiments of the present disclosure. Any neural network that can realize the expression recognition function through training and the neural network that realizes the smile value detection function through training are both It can be applied to the embodiments of the present disclosure. In a possible implementation manner, facial expression detection and smile value detection can also be realized by detecting the key points of the face and the mouth of the target object in the video.
具体在表情检测与微笑值检测达到何种检测结果的情况下，确定目标对象产生目标情绪，其实现方式可以根据实际情况灵活设定。在一种可能的实现方式中，可以认为检测到视频帧中目标对象展示至少一种第一目标表情，或是微笑值检测的结果超过目标微笑值的情况下，认为该视频帧中的目标对象表现出目标情绪，在这种情况下，可以将该视频帧作为第一检测帧。其中，第一目标表情的具体表情种类可以根据实际情况灵活设定，不局限于下述公开实施例。在一种可能的实现方式中，可以将高兴作为第一目标表情，即可以将检测到的目标对象的表情为高兴的视频帧均作为第一检测帧。在一种可能的实现方式中，可以将高兴与平静均作为第一目标表情，即可以将检测到的目标对象的表情为高兴或平静的视频帧，均作为第一检测帧。同理，目标微笑值的具体数值同样可以根据实际情况进行灵活设定，在此不做具体限定。因此，在一种可能的实现方式中，还可以将微笑值的检测结果超过目标微笑值的视频帧，作为第一检测帧。Specifically, in the case of what kind of detection result is achieved by the expression detection and the smile value detection, it is determined that the target object produces the target emotion, and the realization method can be flexibly set according to the actual situation. In a possible implementation manner, it can be considered that the target object in the video frame is detected to show at least one first target expression, or the smile value detection result exceeds the target smile value, the target object in the video frame is considered Show the target emotion, in this case, the video frame can be used as the first detection frame. Among them, the specific expression type of the first target expression can be flexibly set according to the actual situation, and is not limited to the following disclosed embodiments. In a possible implementation manner, happiness may be used as the first target expression, that is, video frames in which the detected expression of the target object is happy may be used as the first detection frame. In a possible implementation manner, both happy and calm can be used as the first target expression, that is, the detected expression of the target object can be a happy or calm video frame, and both can be used as the first detection frame. In the same way, the specific value of the target smile value can also be flexibly set according to the actual situation, and there is no specific limitation here. Therefore, in a possible implementation manner, a video frame whose smile value detection result exceeds the target smile value may also be used as the first detection frame.
在一种可能的实现方式中，可以在检测到某一视频帧为第一检测帧的情况下，确定目标对象产生目标情绪。在一种可能的实现方式中，为了提高检测的准确性，减小检测误差对学习行为检测结果的影响，可以在检测到连续的第一检测帧的数量超过第三阈值的情况下，确定目标对象产生目标情绪。其中，可以将连续视频帧中每一帧均为第一检测帧的视频帧序列，作为连续的第一检测帧。第三阈值的数量可以为根据实际情况灵活设定的数量，其数值可以与第一阈值或第二阈值相同，也可以不同，在一个示例中，第三阈值的数量可以为6，即检测到连续6帧均为第一检测帧的情况下，可以认为目标对象产生目标情绪。In a possible implementation manner, in the case that a certain video frame is detected as the first detection frame, it may be determined that the target object generates the target emotion. In a possible implementation, in order to improve the accuracy of detection and reduce the impact of detection errors on the results of learning behavior detection, the target can be determined when the number of consecutive first detection frames exceeds the third threshold. The subject develops the target emotion. Wherein, a video frame sequence in which each frame in the continuous video frames is the first detection frame may be used as the continuous first detection frame. The number of the third threshold can be a number flexibly set according to the actual situation, and its value can be the same as or different from the first threshold or the second threshold. In an example, the number of the third threshold can be 6, which means it is detected In the case where 6 consecutive frames are the first detection frame, it can be considered that the target object has the target emotion.
进一步地，在确定目标对象产生目标情绪以后，还可以从连续的第一检测帧中选定一帧作为目标情绪开始帧，然后在目标情绪开始帧以后，连续10帧未检测到目标对象的表情为第一目标表情，或是连续10帧中目标对象的微笑值检测结果不超过第三阈值，或是某帧或某几帧检测不到目标对象的情况下，可以进一步确定目标情绪结束帧，然后根据目标情绪开始帧或是目标情绪结束帧来确定目标对象产生目标情绪的次数和/或时间等，具体的过程可以参考目标手势的相应过程，在此不再赘述。Further, after it is determined that the target object produces the target emotion, a frame from the first continuous detection frame can be selected as the target emotion start frame, and then after the target emotion start frame, the expression of the target object is not detected for 10 consecutive frames If it is the first target expression, or the smile value detection result of the target object in 10 consecutive frames does not exceed the third threshold, or the target object cannot be detected in a certain frame or a few frames, the target emotion end frame can be further determined, Then, according to the target emotion start frame or the target emotion end frame, the number and/or time of the target emotion generated by the target object are determined. The specific process can refer to the corresponding process of the target gesture, which will not be repeated here.
通过对包含目标对象的视频帧进行表情检测和/或微笑值检测，并根据表情检测以及微笑值检测的结果，来确定第一检测帧，从而在检测到连续的第一检测帧的数量超过第三阈值的情况下，确定目标对象产生目标情绪，通过上述过程，可以基于目标对象的表情以及微笑程度来灵活确定目标对象在学习过程中的情绪，从而可以更加全面和准确地感知目标对象在学习过程中的情绪状态，生成更为准确的学习状态信息。By performing expression detection and/or smile value detection on the video frame containing the target object, and according to the results of the expression detection and smile value detection, the first detection frame is determined, so that when the number of consecutive first detection frames exceeds the first detection frame In the case of three thresholds, it is determined that the target object produces the target emotion. Through the above process, the emotion of the target object in the learning process can be flexibly determined based on the expression and smile of the target object, so that the target object can be more comprehensively and accurately perceived in the learning process. The emotional state in the process generates more accurate learning state information.
在一种可能的实现方式中，学习行为可以包括：关注教学课程的展示区域；In a possible implementation, the learning behavior can include: paying attention to the display area of the teaching course;
在这种情况下，对包含目标对象的视频帧进行至少一类学习行为检测，可以包括：In this case, performing at least one type of learning behavior detection on the video frame containing the target object may include:
对包含目标对象的视频帧进行表情检测和人脸角度检测；Perform expression detection and face angle detection on video frames containing target objects;
在检测到视频帧中目标对象展示至少一种第二目标表情且人脸角度在目标人脸角度范围以内的情况下，将检测到的视频帧作为第二检测帧；In a case where it is detected that the target object in the video frame shows at least one second target expression and the face angle is within the target face angle range, use the detected video frame as the second detection frame;
在检测到连续的第二检测帧的数量超过第四阈值的情况下，确定目标对象关注教学课程的展示区域。When it is detected that the number of consecutive second detection frames exceeds the fourth threshold, it is determined that the target object pays attention to the display area of the teaching course.
其中，教学课程的展示区域的实现形式可以参考上述各公开实施例，在此不再赘述。Among them, the implementation form of the display area of the teaching course can refer to the above-mentioned disclosed embodiments, which will not be repeated here.
通过上述公开实施例可以看出，在学习行为包括关注教学课程的展示区域的情况下，可以通过表情检测和人脸角度检测，来实现目标对象的学习行为检测。在一种可能的实现方式中，也可以仅通过人脸角度检测来实现关注教学课程的展示区域这一学习行为的检测。后续各公开实施例均以通过表情检测与人脸角度检测来确定目标对象是否关注教学课程的展示区域为例进行说明，其余实现方式可以参考后续各公开实施例进行相应扩展，在此不再赘述。It can be seen from the above disclosed embodiments that in the case where the learning behavior includes paying attention to the display area of the teaching course, the learning behavior detection of the target object can be achieved through expression detection and face angle detection. In a possible implementation manner, the detection of the learning behavior of paying attention to the display area of the teaching course can also be realized only by detecting the face angle. Subsequent disclosed embodiments are described by using expression detection and face angle detection to determine whether the target object pays attention to the display area of the teaching course as an example. The remaining implementation methods can be expanded with reference to the subsequent disclosed embodiments, and will not be repeated here. .
其中，表情检测的实现方式可以参考上述各公开实施例，在此不再赘述；人脸角度检测可以是对人脸的朝向角度等进行检测。具体的人脸角度检测方式可以根据实际情况灵活决定，任何能检测到目标对象的人脸角度的方式，均可以作为人脸角度检测的检测方式，不局限于下述各公开实施例。在一种可能的实现方式中，可以通过人脸角度检测神经网络，来实现目标对象的人脸角度检测。具体地 人脸角度检测神经网络的结构与实现方式在本公开实施例中不做限定，任何可以通过训练实现人脸角度检测功能的神经网络均可以应用于本公开实施例。在一种可能的实现方式中，也可以通过对视频中目标对象的人脸关键点进行检测，来确定目标对象的人脸角度。人脸角度检测可以检测出的人脸的角度的形式也可以根据实际情况灵活决定，在一种可能的实现方式中，可以通过检测出目标对象的人脸的偏航角与俯仰角，来确定目标对象的人脸角度。Among them, the implementation of expression detection can refer to the above disclosed embodiments, which will not be repeated here; the face angle detection can be the detection of the orientation angle of the face. The specific face angle detection method can be flexibly determined according to the actual situation. Any method that can detect the face angle of the target object can be used as the face angle detection method, and is not limited to the following disclosed embodiments. In a possible implementation manner, the face angle detection of the target object can be realized through the face angle detection neural network. Specifically, the structure and implementation of the face angle detection neural network are not limited in the embodiments of the present disclosure, and any neural network that can realize the face angle detection function through training can be applied to the embodiments of the present disclosure. In a possible implementation manner, the face angle of the target object can also be determined by detecting the key points of the target object's face in the video. The form of the face angle that can be detected by the face angle detection can also be flexibly determined according to the actual situation. In a possible implementation, it can be determined by detecting the yaw angle and pitch angle of the target object’s face The angle of the target's face.
具体在表情检测与人脸角度检测达到何种检测结果的情况下，确定目标对象关注教学课程的展示区域，其实现方式可以根据实际情况灵活设定。在一种可能的实现方式中，可以认为检测到视频帧中目标对象展示至少一种第二目标表情，且检测到的人脸角度在目标人脸角度范围以内的情况下，认为该视频帧中的目标对象关注了教学课程的展示区域，在这种情况下，可以将该视频帧作为第二检测帧。其中，第二目标表情的具体表情种类可以根据实际情况灵活设定，可以与上述公开实施例中提到的第一目标表情相同，也可以与上述公开实施例中提到的第一目标表情不同，不局限于下述公开实施例。在一种可能的实现方式中，可以将平静作为第二目标表情，即可以将检测到的目标对象的表情为平静且人脸角度在目标人脸角度范围以内的视频帧均作为第二检测帧。在一种可能的实现方式中，可以将其他以外的表情均作为第二目标表情，即可以将检测到的目标对象的人脸角度在目标人脸角度范围以内，且表情不是“其他”的视频帧，均作为第二检测帧。同理，目标人脸角度范围的具体范围数值同样可以根据实际情况进行灵活设定，在此不做具体限定。在一种可能的实现方式中，该目标人脸角度范围可以是静态的，在一个示例中，可以将教师授课中可能移动到的总体位置(比如线下场景中教师所处的讲台区域等)作为目标人脸角度范围；在一个示例中，可以将目标对象观看教学课程过程中可能关注到的固定区域(比如线上场景中目标对象所关注的显示屏等)作为目标人脸角度范围。在一种可能的实现方式中，该目标人脸角度范围也可以是动态的，在一个示例中，可以根据教师授课中移动的当前位置来灵活确定目标人脸角度范围，即可以随着教师的移动，来动态更改目标人脸角度范围的数值。Specifically, in the case of the detection results achieved by the expression detection and the face angle detection, the target object is determined to focus on the display area of the teaching course, and the implementation method can be flexibly set according to the actual situation. In a possible implementation manner, it can be considered that when it is detected that the target object in the video frame shows at least one second target expression, and the detected face angle is within the range of the target face angle, it can be considered that the video frame The target object of is concerned with the display area of the teaching course. In this case, the video frame can be used as the second detection frame. Among them, the specific expression type of the second target expression can be flexibly set according to the actual situation, and may be the same as the first target expression mentioned in the above-mentioned public embodiment, or it may be different from the first target expression mentioned in the above-mentioned public embodiment It is not limited to the following disclosed embodiments. In a possible implementation manner, calm can be used as the second target expression, that is, the detected target object's expression is calm and the video frame whose face angle is within the range of the target face angle can be regarded as the second detection frame . In a possible implementation manner, all expressions other than other expressions can be used as the second target expression, that is, the face angle of the detected target object can be within the range of the target face angle, and the expression is not "other" video Frames are regarded as the second detection frame. In the same way, the specific range value of the target face angle range can also be flexibly set according to the actual situation, and no specific limitation is made here. In a possible implementation, the target face angle range may be static. In one example, the overall position that the teacher may move to during the lecture (such as the podium area where the teacher is located in the offline scene, etc.) As the target face angle range; in one example, a fixed area (such as the display screen that the target object pays attention to in an online scene) may be taken as the target face angle range when the target object views the teaching course. In a possible implementation, the target face angle range can also be dynamic. In one example, the target face angle range can be flexibly determined according to the current position of the teacher's movement during the lecture, that is, it can follow the teacher's movement To dynamically change the value of the target face angle range.
在一种可能的实现方式中，可以在检测到某一视频帧为第二检测帧的情况下，确定目标对象关注教学课程的展示区域。在一种可能的实现方式中，为了提高检测的准确性，减小检测误差对学习行为检测结果的影响，可以在检测到连续的第二检测帧的数量超过第四阈值的情况下，确定目标对象关注教学课程的展示区域。其中，可以将连续视频帧中每一帧均为第二检测帧的视频帧序列，作为连续的第二检测帧。第四阈值的数量可以为根据实际情况灵活设定的数量，其数值可以与第一阈值、第二阈值或第三阈值相同，也可以不同，在一个示例中，第四阈值的数量可以为6，即检测到连续6帧均为第二检测帧的情况下，可以认为目标对象关注教学课程的展示区域。In a possible implementation manner, in the case that a certain video frame is detected as the second detection frame, it can be determined that the target object pays attention to the display area of the teaching course. In a possible implementation manner, in order to improve the accuracy of detection and reduce the impact of detection errors on the results of learning behavior detection, the target can be determined when the number of consecutive second detection frames exceeds the fourth threshold. The subject pays attention to the display area of the teaching course. Wherein, a video frame sequence in which each frame in the continuous video frames is the second detection frame may be used as the continuous second detection frame. The number of the fourth threshold can be a number flexibly set according to actual conditions, and its value can be the same as or different from the first threshold, the second threshold, or the third threshold. In an example, the number of the fourth threshold can be 6. , That is, when it is detected that 6 consecutive frames are all the second detection frames, it can be considered that the target object pays attention to the display area of the teaching course.
进一步地，在确定目标对象关注教学课程的展示区域以后，还可以从连续的第二检测帧中选定一帧作为关注开始帧，然后在关注开始帧以后，连续10帧未检测到目标对象的表情为第二目标表情，或是连续10帧中目标对象的人脸角度不在目标人脸角度范围以内，或是某帧或某几帧检测不到目标对象的情况下，可以进一步确定关注结束帧，然后根据关注开始帧或是关注结束帧来确定目标对象关注教学课程展示区域的次数和/或时间等，具体的过程可以参考目标手势以及目标情绪的相应过程，在此不再赘述。Further, after determining the display area of the target object's attention to the teaching course, you can also select a frame from the consecutive second detection frames as the attention start frame, and then after the attention start frame, the target object is not detected for 10 consecutive frames When the expression is the second target expression, or the face angle of the target object in 10 consecutive frames is not within the range of the target face angle, or the target object cannot be detected in a certain frame or a few frames, the end frame of attention can be further determined , And then determine the number and/or time the target object pays attention to the teaching course display area according to the focus start frame or focus end frame. The specific process can refer to the corresponding process of target gestures and target emotions, which will not be repeated here.
通过对包含目标对象的视频帧进行表情检测和人脸角度检测，并根据表情检测以及人脸角度检测的结果，来确定第二检测帧，从而在检测到连续的第二检测帧的数量超过第四阈值的情况下，确定目标对象关注教学课程的展示区域，通过上述过程，可以基于目标对象的表情以及人脸角度来灵活确定目标对象是否关注教学课程的展示区域，从而可以更加全面和准确地感知目标对象在学习过程中的精力集中情况，生成更为准确的学习状态信息。By performing expression detection and face angle detection on the video frame containing the target object, and according to the results of expression detection and face angle detection, the second detection frame is determined, so that the number of consecutive second detection frames exceeds the first detection frame. In the case of four thresholds, it is determined that the target object pays attention to the display area of the teaching course. Through the above process, it can be flexibly determined whether the target object pays attention to the display area of the teaching course based on the expression and face angle of the target object, so that it can be more comprehensive and accurate Perceive the energy concentration of the target object in the learning process, and generate more accurate learning status information.
在一种可能的实现方式中，学习行为还可以包括：与其他对象产生至少一种互动行为。互动行为的实现方式可以参考上述各公开实施例，在此不再赘述。在这种情况下，对包含目标对象的视频帧进行互动行为检测的方式可以根据实际情况灵活决定，在一种可能的实现方式中，如果互动行为为线上的互动行为，比如收到老师通过线上课堂发送的小红花，或是根据老师在线上课堂的点名进行发言的情况下，则对互动行为的检测方式可以为直接根据其他对象传递的信号，确定目标对象是否产生互 动行为。在一种可能的实现方式中，如果互动行为为线下的互动行为，比如目标对象在教室中受到老师的点名而发言的情况下，检测目标对象是否发生互动行为的方式可以包括：通过对目标对象的目标动作进行识别，来确定目标对象是否发生互动行为，其中，目标动作可以根据互动行为的实际情况灵活设定，比如目标动作可以包括有起立后发言、或是人脸朝向其他对象且发言时间超过一定时间数值等。In a possible implementation manner, the learning behavior may also include: generating at least one interaction behavior with other objects. For the implementation of the interactive behavior, reference may be made to the above disclosed embodiments, which will not be repeated here. In this case, the method of detecting the interactive behavior of the video frame containing the target object can be flexibly determined according to the actual situation. In a possible implementation, if the interactive behavior is an online interactive behavior, such as receiving the teacher’s approval In the case of the small red flowers sent by the online classroom, or when the teacher speaks according to the roll call of the teacher in the online classroom, the interactive behavior detection method can be directly based on the signals transmitted by other objects to determine whether the target object has an interactive behavior. In a possible implementation, if the interactive behavior is offline, for example, when the target object is called by the teacher to speak in the classroom, the method of detecting whether the target object has an interactive behavior can include: The target action of the object is recognized to determine whether the target object has an interactive behavior. The target action can be flexibly set according to the actual situation of the interactive behavior. For example, the target action can include speaking after standing up or speaking with the face facing other objects. The time exceeds a certain time value, etc.
在一种可能的实现方式中，学习行为还可以包括在视频中的至少部分视频帧中未出现，在这种情况下，步骤S12可以包括：In a possible implementation manner, the learning behavior may also include not appearing in at least part of the video frames in the video. In this case, step S12 may include:
对视频进行目标对象检测，得到包含目标对象的视频帧，并将视频中包含目标对象的视频帧以外的视频帧，作为未检测到目标对象的视频帧；Perform target object detection on the video to obtain a video frame containing the target object, and use the video frame other than the video frame containing the target object as the video frame where the target object is not detected;
在未检测到目标对象的视频帧的数量超过预设视频帧数量的情况下，检测到学习行为包括：在视频中的至少部分视频帧中未出现。In a case where the number of video frames in which the target object is not detected exceeds the preset number of video frames, detecting the learning behavior includes: not appearing in at least part of the video frames in the video.
其中，对视频进行目标对象检测的方式详见上述各公开实施例，在此不再赘述。在一种可能的实现方式中，视频中的各视频帧除了包含目标对象的视频帧以外，还可能存在不包含目标对象的视频帧，因此可以将这些不包含目标对象的视频帧作为未检测到目标对象的视频帧，并在未检测到目标对象的视频帧的数量超过预设视频帧数量的情况下，确认检测到“在视频中的至少部分视频帧中未出现”这一学习行为。预设视频帧数量可以根据实际情况灵活设定，在一种可能的实现方式中，可以将预设视频帧数量设为0，即在视频中包含未检测到目标对象的视频帧的情况下，即认为检测到在视频中的至少部分视频帧中未出现这一学习行为，在一种可能的实现方式中，预设视频帧数量也可以为大于0的数量，具体如何设定可以根据实际情况灵活决定。Among them, the method of performing target object detection on the video is detailed in the above-mentioned disclosed embodiments, and will not be repeated here. In a possible implementation manner, in addition to the video frames that contain the target object, each video frame in the video may also contain video frames that do not contain the target object. Therefore, these video frames that do not contain the target object can be regarded as undetected Video frames of the target object, and in the case where the number of video frames where the target object is not detected exceeds the preset number of video frames, it is confirmed that the learning behavior of "not appearing in at least part of the video frames in the video" is detected. The number of preset video frames can be flexibly set according to the actual situation. In a possible implementation, the number of preset video frames can be set to 0, that is, when the video contains video frames where the target object is not detected, That is to say, it is considered that this learning behavior is not detected in at least part of the video frames in the video. In a possible implementation, the preset number of video frames can also be a number greater than 0. The specific setting can be based on the actual situation. Flexible decision.
在一种可能的实现方式中，学习行为还可以包括闭眼，在这种情况下的学习行为检测方式可以为闭眼检测，闭眼检测的具体过程可以根据实际情况灵活设定，在一个示例中，可以通过具有闭眼检测功能的神经网络来实现，在一个示例中，也可以通过对眼睛及眼球内的关键点检测来确定目标对象是否闭眼等，比如，在检测到眼球内的关键点的情况下，确定目标对象睁眼；在仅检测到眼睛关键点，未检测到眼球内的关键点情况下，确定目标对象闭眼。在一种可能的实现方式中，学习行为还可以包括在教学课程的展示区域内的目光交流，在这种情况下的学习行为检测方式可以参考上述公开实施例中的关注教学课程的展示区域的过程，具体的检测方式可以灵活发生变化，比如可以对目标对象同时进行闭眼与人脸角度检测，将人脸角度在目标人脸角度范围内且无闭眼的视频帧作为第三检测帧，然后在第三检测帧的数量超过某一设定阈值的情况下，认定目标对象在教学课程的展示区域内进行目光交流等。In a possible implementation, the learning behavior can also include closed eyes. In this case, the learning behavior detection method can be closed eyes detection. The specific process of closed eyes detection can be flexibly set according to the actual situation. In an example , It can be realized by a neural network with closed eyes detection function. In one example, it can also determine whether the target object has closed eyes or not by detecting the key points in the eyes and the eyeball. For example, after detecting the key points in the eyeball In the case of dots, it is determined that the target object has eyes open; in the case that only the key points of the eye are detected, and the key points in the eyeball are not detected, the eyes of the target object are determined to be closed. In a possible implementation manner, the learning behavior can also include eye contact in the display area of the teaching course. In this case, the learning behavior detection method can refer to the focus on the display area of the teaching course in the above disclosed embodiment. During the process, the specific detection method can be flexibly changed. For example, the target object can be detected with closed eyes and face angle at the same time, and the video frame with the face angle within the target face angle range without closed eyes is used as the third detection frame. Then, when the number of third detection frames exceeds a certain set threshold, it is determined that the target object is making eye contact in the display area of the teaching course.
在通过上述公开实施例的各种实现方式的任意组合，以实现对目标对象的至少一类学习行为的检测以后，可以在检测到目标对象执行至少一类学习行为的情况下，通过步骤S13生成学习状态信息。步骤S13的具体实现方式不受限定，可以根据检测到的学习行为的实际情况所灵活变化，不局限于下述各公开实施例。After the detection of at least one type of learning behavior of the target object is achieved through any combination of the various implementation manners of the above disclosed embodiments, it can be generated through step S13 when the target object is detected to perform at least one type of learning behavior. Learning status information. The specific implementation of step S13 is not limited, and can be flexibly changed according to the actual situation of the detected learning behavior, and is not limited to the following disclosed embodiments.
通过上述公开实施例中步骤S13的实际内容可以看出，步骤S13在生成学习状态信息的过程中，可能存在如下几种生成方式，比如可以根据包含至少一类学习行为的视频帧来生成学习状态信息；或是根据目标对象执行至少一类学习行为的持续时间来生成学习状态信息；或是对上述两种情况进行组合，既根据包含至少一类学习行为的视频帧来生成一部分学习状态信息，又根据目标对象执行至少一类学习行为的持续时间来生成另外一类学习状态信息。在既可以根据学习行为的视频帧来生成学习状态信息，又可以根据目标对象执行至少一类学习行为的持续时间来生成学习状态信息的情况下，具体根据哪类学习行为对应生成哪种学习状态信息，其映射方式可以根据实际情况灵活设定。在一种可能的实现方式中，可以将一些积极的学习行为与根据包含学习行为的视频帧来生成学习状态信息这一过程相对应，比如在目标对象执行至少一种目标手势、展现积极的目标情绪、关注教学课程的展示区域以及与其他对象产生至少一种互动行为等情况下，可以根据包含上述学习行为的视频帧，来生成学习状态信息；在一种可能的实现方式中，也可以将一些消极的学习行为，比如目标对象在视频中的至少部分视频帧中未出现、闭眼或是在教学课程的展示区域内未进行目光交流等情况下，可以根据上述学 习行为的持续时间，来生成学习状态信息。From the actual content of step S13 in the above disclosed embodiment, it can be seen that in the process of generating learning state information in step S13, there may be the following generation methods. For example, the learning state can be generated based on a video frame containing at least one type of learning behavior. Information; or generate learning state information according to the duration of the target object performing at least one type of learning behavior; or a combination of the above two situations, both based on the video frame containing at least one type of learning behavior to generate part of the learning state information, Another type of learning state information is generated according to the duration of at least one type of learning behavior performed by the target object. When the learning state information can be generated based on the video frames of the learning behavior, and the learning state information can be generated based on the duration of the target object performing at least one type of learning behavior, which learning state is generated according to which type of learning behavior Information and its mapping method can be flexibly set according to the actual situation. In a possible implementation, some positive learning behaviors can be corresponded to the process of generating learning state information based on the video frames containing the learning behaviors, such as performing at least one target gesture on the target object and showing a positive goal. In the case of emotions, paying attention to the display area of the teaching course, and at least one interactive behavior with other objects, the learning state information can be generated based on the video frame containing the above learning behavior; in a possible implementation manner, it can also be Some negative learning behaviors, such as when the target object does not appear in at least part of the video frame in the video, eyes are closed, or there is no eye contact in the display area of the teaching course, can be based on the duration of the above learning behavior. Generate learning status information.
在一种可能的实现方式中，根据至少部分包含至少一类学习行为的视频帧，生成学习状态信息，可以包括：In a possible implementation manner, generating learning state information according to video frames containing at least one type of learning behavior at least in part may include:
步骤S1311，获取视频中包含至少一类学习行为的视频帧，作为目标视频帧集合；Step S1311: Obtain video frames containing at least one type of learning behavior in the video as a target video frame set;
步骤S1312，对目标视频帧集合中的至少一个视频帧进行人脸质量检测，将人脸质量大于人脸质量阈值的视频帧作为目标视频帧；Step S1312: Perform face quality detection on at least one video frame in the target video frame set, and use a video frame with a face quality greater than a face quality threshold as a target video frame;
步骤S1313，根据目标视频帧，生成学习状态信息。Step S1313: Generate learning state information according to the target video frame.
其中，包含至少一类学习行为的视频帧，可以是在学习行为检测的过程中，检测到目标对象执行其中至少一类行为的视频帧，比如上述公开实施例中提到的第一检测帧、第二检测帧以及第三检测帧等，或是在手势开始帧与手势结束帧之间的包含目标手势的视频帧等。Wherein, the video frame containing at least one type of learning behavior may be a video frame in which the target object is detected to perform at least one type of behavior in the process of learning behavior detection, such as the first detection frame mentioned in the above-mentioned disclosed embodiment, The second detection frame and the third detection frame, etc., or the video frame containing the target gesture between the gesture start frame and the gesture end frame, etc.
在确定了包含至少一类学习行为的视频帧以后，如何得到目标视频帧集合，其实现方式可以灵活决定。在一种可能的实现方式中，可以按照学习行为的类别，分别获取包含每类学习行为的每个视频帧，从而组成每类学习行为的目标视频帧集合；在一种可能的实现方式中，也可以按照学习行为的类别，分别获取包含每类学习行为的部分帧等，然后基于每类学习行为的部分帧来得到该类学习行为的目标视频帧集合，具体选择哪些部分帧，其选择方式可以灵活决定。After the video frames containing at least one type of learning behavior are determined, how to obtain the target video frame set can be flexibly determined. In a possible implementation manner, each video frame containing each type of learning behavior can be obtained according to the type of learning behavior, so as to form the target video frame set of each type of learning behavior; in a possible implementation manner, It is also possible to obtain partial frames containing each type of learning behavior according to the type of learning behavior, and then obtain the target video frame set of that type of learning behavior based on the partial frames of each type of learning behavior, which part of the frame is specifically selected, and the selection method Can be flexibly decided.
在得到了与学习行为对应的目标视频帧集合以后，可以通过步骤S1312，来从目标视频帧集合中选择得到目标视频帧。通过步骤S1312可以看出，在一种可能的实现方式中，可以对目标视频帧集合中的视频帧进行人脸质量检测，然后将人脸质量大于人脸质量阈值的视频帧作为目标视频帧。After obtaining the target video frame set corresponding to the learning behavior, step S1312 may be used to select and obtain the target video frame from the target video frame set. It can be seen from step S1312 that, in a possible implementation manner, face quality detection may be performed on the video frames in the target video frame set, and then video frames with face quality greater than the face quality threshold are used as the target video frames.
其中，人脸质量的检测方式可以根据实际情况灵活设定，不局限于下述公开实施例，在一种可能的实现方式中，可以通过对视频帧中的人脸进行人脸识别，从而确定视频帧中人脸的完整度来确定人脸质量；在一种可能的实现方式中，也可以基于视频帧中人脸的清晰度来确定人脸质量；在一种可能的实现方式中，也可以基于视频帧人脸的完整度、清晰度以及亮度等多个参数来综合评判视频帧中的人脸质量；在一种可能的实现方式中，还可以通过将视频帧输入到人脸质量神经网络，来得到视频帧中的人脸质量，人脸质量神经网络可以通过大量包含人脸质量打分标注的人脸图片训练得到，其具体实现形式可以根据实际情况灵活选择，在本公开实施例中不做限制。Among them, the face quality detection method can be flexibly set according to the actual situation, and is not limited to the following disclosed embodiments. In a possible implementation manner, the face quality can be determined by performing face recognition on the face in the video frame. The completeness of the face in the video frame is used to determine the face quality; in a possible implementation, the face quality can also be determined based on the clarity of the face in the video frame; in a possible implementation, it is also The face quality in the video frame can be comprehensively judged based on multiple parameters such as the completeness, clarity, and brightness of the face of the video frame; in a possible implementation, the video frame can also be input to the face quality nerve Network to obtain the face quality in the video frame. The face quality neural network can be obtained by training a large number of face images containing face quality scores. The specific implementation form can be flexibly selected according to the actual situation. In the embodiments of the present disclosure, No restrictions.
人脸质量阈值的具体数值可以根据实际情况灵活决定，本公开实施例对此不做限制。在一种可能的实现方式中，可以分别为每类学习行为设置不同的人脸质量阈值；在一种可能的实现方式中，也可以分别为每类学习行为设置相同的人脸阈值。在一种可能的实现方式中，还可以将人脸质量阈值设置为目标视频帧集合中人脸质量的最大值，在这种情况下，可以直接将每类学习行为下，人脸质量最高的视频帧作为目标视频帧。The specific value of the face quality threshold can be flexibly determined according to the actual situation, which is not limited in the embodiment of the present disclosure. In a possible implementation manner, different face quality thresholds may be set for each type of learning behavior; in a possible implementation manner, the same face threshold may also be set for each type of learning behavior. In a possible implementation, the face quality threshold can also be set to the maximum value of the face quality in the target video frame set. In this case, you can directly set the highest face quality under each type of learning behavior The video frame is used as the target video frame.
在一些可能的实现方式中，可能存在某些视频帧，同时包含多类学习行为，在这种情况下，处理包含多类学习行为的视频帧的方式可以根据实际情况灵活变化。在一种可能的实现方式中，可以将这些视频帧分别归属在每类学习行为下，然后从每类学习行为对应的视频帧集合中按照步骤S1312进行选择，来得到目标视频帧；在一种可能的实现方式中，也可以直接将同时包含多类学习行为的视频帧选定为目标视频帧。In some possible implementation manners, there may be certain video frames that contain multiple types of learning behaviors at the same time. In this case, the manner of processing video frames containing multiple types of learning behaviors can be flexibly changed according to actual conditions. In a possible implementation manner, these video frames can be assigned to each type of learning behavior, and then selected from the set of video frames corresponding to each type of learning behavior in step S1312 to obtain the target video frame; In a possible implementation manner, a video frame containing multiple types of learning behaviors at the same time can also be directly selected as the target video frame.
在通过上述任意实施例确定目标视频帧以后，可以通过步骤S1313，来根据目标视频帧生成学习状态信息。步骤S1313的实现方式可以根据实际情况灵活选择，详见下述各公开实施例，在此先不做展开。After the target video frame is determined through any of the foregoing embodiments, step S1313 may be used to generate learning state information according to the target video frame. The implementation of step S1313 can be flexibly selected according to the actual situation. For details, please refer to the following disclosed embodiments, which will not be expanded here.
在本公开实施例中，通过获取视频帧中包含至少一类学习行为的视频帧，作为目标视频帧集合，从而根据每类学习行为的目标视频帧集合，选定人脸质量较高的视频帧作为目标视频帧，继而根据目标视频帧来生成学习状态信息。通过上述过程，可以使得生成的学习状态信息，是基于具有较高人脸质量且包含有学习行为的视频帧所得到的信息，具有更高的准确性，从而可以更加精准地把握目标对象的学习状态。In the embodiment of the present disclosure, the video frame containing at least one type of learning behavior is obtained as the target video frame set, so that according to the target video frame set of each type of learning behavior, the video frame with higher face quality is selected As the target video frame, the learning state information is then generated according to the target video frame. Through the above process, the generated learning status information can be based on the information obtained from the video frames with higher face quality and containing learning behaviors, with higher accuracy, so that the learning of the target object can be grasped more accurately state.
如上述公开实施例所述，步骤S1313的实现方式可以灵活变化。在一种可能的实现方式中，步 骤S1313可以包括：As described in the above disclosed embodiment, the implementation of step S1313 can be flexibly changed. In a possible implementation manner, step S1313 may include:
将目标视频帧中的至少一帧作为学习状态信息；和/或，Use at least one of the target video frames as learning state information; and/or,
识别在至少一帧目标视频帧中目标对象所在区域，基于目标对象所在区域，生成学习状态信息。Identify the area where the target object is located in at least one frame of the target video frame, and generate learning state information based on the area where the target object is located.
通过上述公开实施例可以看出，在一种可能的实现方式中，可以直接将目标视频帧中的至少一帧作为学习状态信息，在一个示例中，可以对得到的目标视频帧进行进一步的选定，这一选定可以是随机的，也可以是有一定条件的，然后将选定的目标视频帧直接作为学习状态信息；在一个示例中，也可以直接将得到的每个目标视频帧均作为学习状态信息。It can be seen from the above disclosed embodiments that in a possible implementation manner, at least one frame of the target video frame can be directly used as the learning state information. In an example, the obtained target video frame can be further selected. This selection can be random or subject to certain conditions, and then the selected target video frame is directly used as the learning state information; in one example, each target video frame obtained can also be directly equalized. As learning status information.
在一种可能的实现方式中，还可以对目标视频帧中的目标对象所在区域进行进一步识别，从而根据目标对象所在的区域来生成学习状态信息。其中，识别目标对象区域的方式在本公开实施例中不做限定，在一种可能的实现方式中，可以通过上述公开实施例中提到的具有目标对象检测功能的神经网络来实现。在确定了目标对象在目标视频帧中的区域后，可以进一步对目标视频帧进行相应处理，来得到学习状态信息。其中，处理的方式可以灵活决定，在一个示例中，可以将目标视频帧中目标对象所在区域的图像，作为学习状态信息；在一个示例中，也可以对目标视频帧目标对象所在区域以外的背景区域进行渲染，比如增加其他贴纸，或是对背景区域增加马赛克，或是替换背景区域的图像等，来得到不显示目标对象当前背景的学习状态信息，从而可以对目标对象进行更好的隐私保护，也可以利用贴纸等渲染方式，增加学习状态信息的多样性和美观。In a possible implementation manner, the area where the target object is located in the target video frame may be further identified, so as to generate learning state information according to the area where the target object is located. The method of recognizing the target object area is not limited in the embodiment of the present disclosure. In a possible implementation manner, it can be implemented by the neural network with the target object detection function mentioned in the above-mentioned disclosed embodiment. After the area of the target object in the target video frame is determined, the target video frame can be further processed accordingly to obtain the learning state information. Among them, the processing method can be flexibly determined. In one example, the image of the area where the target object is located in the target video frame can be used as the learning state information; in one example, the background outside the area where the target object is located in the target video frame can also be used as learning state information. Area rendering, such as adding other stickers, or adding mosaic to the background area, or replacing the image of the background area, etc., to get the learning status information that does not display the current background of the target object, so as to better protect the privacy of the target object , You can also use stickers and other rendering methods to increase the diversity and beauty of the learning status information.
通过将目标视频中的至少一帧作为学习状态信息，和/或根据目标视频帧中目标对象所在区域来生成学习状态信息，通过上述方式，可以使得最终得到的学习状态信息更为灵活，从而可以根据目标对象的需求，来得到更加突出目标对象的学习状态信息，或是更为保护目标对象隐私的学习状态信息。By using at least one frame in the target video as the learning state information, and/or generating the learning state information according to the area of the target object in the target video frame, the above method can make the final learning state information more flexible, so that According to the needs of the target object, the learning status information of the target object is more prominent, or the learning status information that protects the privacy of the target object more can be obtained.
上述各公开实施例可以通过任意组合，来得到以包含学习行为的视频帧为基础所生成的学习状态信息，比如表1示出根据本公开一实施例的学习状态信息生成规则。The above disclosed embodiments can be combined arbitrarily to obtain learning state information generated based on video frames containing learning behaviors. For example, Table 1 shows a learning state information generation rule according to an embodiment of the present disclosure.
Figure PCTCN2020137690-appb-000001
Figure PCTCN2020137690-appb-000001
Figure PCTCN2020137690-appb-000002
Figure PCTCN2020137690-appb-000002
表1 学习状态信息生成规则Table 1 Rules for generating learning status information
其中，M、N、X、Y、Z均为正整数，具体数值可根据实际需求来设定。且对于表1中处于不同行的M等参数，可以相同或不同，上述M等参数仅作为一种示意说明，并不作为对本公开内容的限定。Among them, M, N, X, Y, and Z are all positive integers, and the specific values can be set according to actual needs. In addition, the parameters such as M in different rows in Table 1 may be the same or different. The above-mentioned parameters such as M are only used as a schematic description, and not as a limitation to the present disclosure.
其中，精彩时刻为目标对象产生积极学习行为所对应的时刻。通过表1可以看出，在一个示例中，可以在检测到目标对象执行举手等目标手势、产生开心这一目标情绪、或是聚精会神关注教学课程的展示区域以及与老师产生点名发言等互动等学校行为的情况下，对视频进行一定的数据处理，并在数据处理后，对视频帧进行进一步地图像处理，从而得到目标视频帧作为学习状态信息。Among them, the wonderful moment is the moment corresponding to the positive learning behavior of the target object. It can be seen from Table 1 that in an example, the target object can be detected to perform target gestures such as raising hands, to generate the target emotion of happiness, or to pay attention to the display area of the teaching course, and to have a roll call with the teacher. In the case of school behavior, certain data processing is performed on the video, and after the data processing, further image processing is performed on the video frame to obtain the target video frame as the learning state information.
在一种可能的实现方式中，根据目标对象执行至少一类学习行为的持续时间，生成学习状态信息，可以包括：In a possible implementation manner, generating learning state information according to the duration of the target object performing at least one type of learning behavior may include:
步骤S1321，在检测到目标对象执行至少一类学习行为的时间不小于时间阈值的情况下，记录至少一类学习行为的持续时间；Step S1321, in the case where it is detected that the time for the target object to perform at least one type of learning behavior is not less than the time threshold, record the duration of the at least one type of learning behavior;
步骤S1322，将至少一类学习行为对应的持续时间，作为学习状态信息。In step S1322, the duration corresponding to at least one type of learning behavior is used as the learning state information.
其中，时间阈值可以是根据实际情况灵活设定的某一数值，不同类学习行为的时间阈值可以相同，也可以不同。在检测到目标对象在一定时间内执行某一类学习行为的情况下，可以统计目标对象执行这些学习行为的时间，从而作为学习状态信息反馈到老师或家长处。具体的统计条件以及在哪些学习行为下统计时间，其实现方式均可以根据实际情况灵活设定。Among them, the time threshold can be a certain value flexibly set according to the actual situation, and the time thresholds of different types of learning behaviors can be the same or different. When it is detected that the target object performs a certain type of learning behavior within a certain period of time, the time for the target object to perform these learning behaviors can be counted, so as to feed back to the teacher or parent as learning status information. The specific statistical conditions and the statistical time under which learning behaviors can be implemented can be flexibly set according to the actual situation.
在一种可能的实现方式中，在检测到目标对象的未出现在视频中(比如视频中无人、视频帧中有人但无法确定是否为目标对象或是镜头中有人但并非目标对象)的时间超过一定时长、目标对象闭眼或是目标对象未观看教学课程的展示区域的情况下，可以统计这些学习行为的时长并将其作为学习状态信息。In a possible implementation, when it is detected that the target object does not appear in the video (such as no one in the video, someone in the video frame but it is impossible to determine whether it is the target object or there is someone in the shot but not the target object) In the case that the target object has closed eyes or the target object does not watch the display area of the teaching course, the time length of these learning behaviors can be counted and used as the learning status information.
在本公开实施例中，通过在检测到目标对象执行至少一类学习行为的时间不小于时间阈值的情况下，记录至少一类学习行为的持续时间并作为学习状态信息，通过上述过程，可以将学习状态信息进行量化，更为直观且精确地掌握目标对象的学习状态。In the embodiments of the present disclosure, when it is detected that the time for the target object to perform at least one type of learning behavior is not less than the time threshold, the duration of at least one type of learning behavior is recorded as the learning state information. Through the above process, the The learning status information is quantified, and the learning status of the target object can be grasped more intuitively and accurately.
在一种可能的实现方式中，本公开实施例中提出的视频处理方法，还可以包括：In a possible implementation, the video processing method proposed in the embodiment of the present disclosure may further include:
对视频中的至少部分视频帧中的背景区域进行渲染，其中，背景区域为视频帧中目标对象以外的区域。Rendering the background area in at least part of the video frame in the video, where the background area is an area outside the target object in the video frame.
其中，背景区域的分割方式，以及对背景区域的渲染方式，可以参考上述公开实施例中，对目标视频帧中目标对象所在区域进行识别以及识别后的渲染过程，在此不再赘述。对背景区域进行渲染的过程中，在一个示例中，可以通过当前的视频处理装置中预设的通用模板进行渲染；在一个示例中，也可以通过调用非视频处理装置的数据库中的其他模板或定制模板等进行渲染，比如可以从非视频处理装置的云端服务器中，调用其他的背景模板等，对视频中的背景区域进行渲染等。For the segmentation method of the background area and the rendering method of the background area, reference may be made to the above-mentioned disclosed embodiment for identifying the area where the target object in the target video frame is located and the rendering process after the recognition, which will not be repeated here. In the process of rendering the background area, in one example, it can be rendered by a universal template preset in the current video processing device; in one example, it can also be rendered by calling other templates in the database of the non-video processing device or Customized templates, etc. for rendering, for example, other background templates can be called from a cloud server of a non-video processing device, etc., to render the background area in the video, etc.
通过对视频中的至少部分视频帧中的背景区域进行渲染，一方面可以保护视频中目标对象的隐私，减小目标对象由于没有合适的视频采集位置导致隐私泄露的可能性，另一方面，也可以增强目标对象观看教学课程过程的趣味性。By rendering the background area in at least part of the video frame in the video, on the one hand, the privacy of the target object in the video can be protected, and the possibility of privacy leakage of the target object due to the lack of a suitable video capture location is reduced. On the other hand, it is also It can enhance the interest of the target object to watch the teaching course process.
在一种可能的实现方式中，本公开实施例中提出的视频处理方法，还可以包括：In a possible implementation, the video processing method proposed in the embodiment of the present disclosure may further include:
统计至少一个目标对象的学习状态信息，得到至少一个目标对象的统计结果；Calculate the learning status information of at least one target object, and obtain a statistical result of at least one target object;
根据至少一个目标对象的统计结果，生成学习状态统计数据。According to the statistical result of at least one target object, the learning state statistical data is generated.
在本公开实施例中，一个视频中包含的目标对象可以为一个，也可以为多个，另外，本公开实施例中的视频处理方法，可以用于对单个视频进行处理，也可以用于对多个视频进行处理。因此，相应的，可以得到一个目标对象的学习状态信息，也可以得到多个目标对象的学习状态信息。在这种情况下，可以对至少一个目标对象的学习状态信息进行统计，来得到至少一个目标对象的统计结果。其中，统计结果可以包含有目标对象的学习状态信息以外，还可以包含有其他与目标对象观看教学课程所相关的信息。比如，在一种可能的实现方式中，在步骤S12以前，即对目标对象进行学习行为检测以前，还可以获取目标对象的签到数据。目标对象的签到数据可以包含有目标对象的身份信息以及签 到时间等，具体签到数据的获取方式可以根据目标对象的实际签到方式所灵活决定，在本公开实施例中不做限定。In the embodiment of the present disclosure, the target object contained in a video may be one or multiple. In addition, the video processing method in the embodiment of the present disclosure may be used to process a single video, or it may be used to process a single video. Multiple videos are processed. Therefore, correspondingly, the learning status information of one target object can be obtained, and the learning status information of multiple target objects can also be obtained. In this case, statistics can be performed on the learning state information of at least one target object to obtain a statistical result of at least one target object. Among them, the statistical result may include not only the learning status information of the target object, but also other information related to the target object's viewing of the teaching course. For example, in a possible implementation manner, before step S12, that is, before performing learning behavior detection on the target object, the sign-in data of the target object can also be obtained. The check-in data of the target object may include the identity information and check-in time of the target object. The specific check-in data acquisition method can be flexibly determined according to the actual check-in method of the target object, which is not limited in the embodiments of the present disclosure.
在得到了至少一个目标对象的统计结果以后，可以根据至少一个统计结果生成学习状态统计数据。具体地，学习状态统计数据的生成方式与内容，可以根据统计结果的实现形式所灵活变化。详见下述各公开实施例，在此先不做展开。After the statistical result of at least one target object is obtained, the learning state statistical data can be generated according to the at least one statistical result. Specifically, the generation method and content of the learning state statistical data can be flexibly changed according to the realization form of the statistical result. For details, please refer to the following disclosed embodiments, which will not be expanded here.
在本公开实施例中，通过统计至少一个目标对象的学习状态信息，得到至少一个目标对象的统计结果，从而根据至少一个目标对象的统计结果来生成学习状态统计数据，通过上述过程，可以有效地对多个目标对象的学习状态进行综合评估，从而更加便于教师掌握整个课堂的整体学习情况，也便于其他相关人员更加全面的了解目标对象当前所处的学习位置等。In the embodiment of the present disclosure, the statistical result of the at least one target object is obtained by counting the learning status information of at least one target object, so as to generate the learning status statistical data according to the statistical result of the at least one target object. Through the above process, it is possible to effectively Comprehensive evaluation of the learning status of multiple target objects makes it easier for teachers to grasp the overall learning situation of the entire classroom, and it is also convenient for other relevant personnel to have a more comprehensive understanding of the current learning position of the target object.
在一种可能的实现方式中，根据至少一个所述目标对象的统计结果，生成学习状态统计数据，包括：In a possible implementation manner, generating the learning state statistical data according to the statistical result of at least one of the target objects includes:
根据至少一个目标对象所属的类别，获取至少一个类别包含的目标对象的统计结果，生成至少一个类别的学习状态统计数据，其中，目标对象所属的类别包括目标对象参与的课程、目标对象注册的机构以及目标对象使用的设备中的至少一种；和/或，According to the category to which at least one target object belongs, the statistical result of the target object contained in the at least one category is obtained, and the learning status statistical data of at least one category is generated. And at least one of the devices used by the target object; and/or,
将至少一个目标对象的统计结果进行可视化处理，生成至少一个目标对象的学习状态统计数据。Visualizing the statistical results of the at least one target object to generate statistical data of the learning state of the at least one target object.
其中，目标对象所属的类别可以是根据目标对象的身份所划分的类别，举例来说，目标对象所属的类别可以包括目标对象参与的课程、目标对象注册的机构以及目标对象使用的设备中的至少一种，其中，目标对象参与的课程可以是上述公开实施例中提到的目标对象观看的教学课程，目标对象注册的机构可以是目标对象所在的教育机构、或是目标对象所在的年级或是目标对象所在的班级等，目标对象使用的设备可以是线上场景中，目标对象参加在线课程所使用的终端设备等。The category to which the target object belongs may be a category divided according to the identity of the target object. For example, the category to which the target object belongs may include at least one of the courses the target object participates in, the institution registered by the target object, and the equipment used by the target object. In one type, the course that the target object participates in may be the teaching course watched by the target object mentioned in the above disclosed embodiment, and the institution registered by the target object may be the educational institution where the target object is located, or the grade or grade of the target object. The class where the target object is located, and the equipment used by the target object may be the terminal device used by the target object to participate in the online course in an online scene.
在本公开实施例中，可以根据目标对象所属的类别，来获取至少一个类别包含的目标对象的统计结果，即可以将目标对象所属类别下的至少一个统计结果进行汇总，来得到该类别下的整体学习状态统计数据。举例来说，可以按照使用设备、课程、教育机构等类别进行划分，分别得到同一设备下不同目标对象的统计结果、同一课程下不同目标对象的统计结果以及同一教育机构中不同目标对象的统计结果等。在一个示例中，还可以将这些统计结果以报表的形式进行展现。在一个示例中，报表中每个类别下的统计结果，既可以包含有每个目标对象的总体学习状态信息，还可以包含有每个目标对象的具体学习状态信息，比如关注教学课程展示区域的时间长度、微笑的时间长度等，除此以外，还可以包含有其他与观看教学课程相关的信息，比如目标对象的签到时间、签到次数、目标对象和预设数据库中的人脸匹配的情况、签到设备以及签到课程等。In the embodiments of the present disclosure, the statistical results of the target objects contained in at least one category can be obtained according to the category to which the target object belongs, that is, at least one statistical result of the category to which the target object belongs can be summarized to obtain the Statistics of overall learning status. For example, it can be divided according to the categories of equipment, courses, educational institutions, etc., and the statistical results of different target objects under the same equipment, the statistical results of different target objects under the same course, and the statistical results of different target objects in the same educational institution can be obtained respectively. Wait. In an example, these statistical results can also be displayed in the form of a report. In an example, the statistical results of each category in the report can include not only the overall learning status information of each target object, but also the specific learning status information of each target object, such as the focus on the display area of the teaching course The length of time, the length of smiling time, etc., in addition to this, it can also contain other information related to watching the teaching course, such as the check-in time of the target object, the number of check-ins, the match between the target object and the face in the preset database, Sign-in equipment and sign-in courses, etc.
除此之外，还可以对至少一个目标对象的统计结果进行可视化处理，来得到至少一个目标对象的学习状态统计数据。其中，可视化处理的方式可以根据实际情况灵活决定，比如可以将数据整理成图表或视频等形式。学习状态统计数据中包含的内容可以根据实际情况灵活决定，比如可以包含有目标对象的总体学习状态信息、目标对象观看的教学课程名称以及目标对象的具体学习状态信息等，具体包含哪些数据可以根据实际情况灵活设定。在一个示例中，可以将目标对象的身份、目标对象观看的教学课程名称、目标对象的关注教学课程展示区域的时长、目标对象的关注程度强弱、目标对象与其他目标对象之间的数据比较结果、目标对象的互动次数以及目标对象的情绪等内容，整理成可视化的报告，并发送给目标对象或目标对象的其他相关人员，比如目标对象的家长等。In addition, the statistical results of at least one target object can also be visualized to obtain the statistical data of the learning state of the at least one target object. Among them, the visual processing method can be flexibly determined according to the actual situation, for example, the data can be sorted into forms such as charts or videos. The content contained in the learning status statistics can be flexibly determined according to the actual situation. For example, it can include the overall learning status information of the target object, the name of the teaching course watched by the target object, and the specific learning status information of the target object. The actual situation is flexible. In one example, the identity of the target object, the name of the teaching course viewed by the target object, the duration of the display area of the target object’s attention teaching course, the degree of attention of the target object, and the data comparison between the target object and other target objects The results, the number of interactions of the target object, and the emotions of the target object are organized into a visual report, and sent to the target object or other relevant personnel of the target object, such as the parents of the target object.
在一个示例中，可视化处理后的学习状态统计数据除了图片与视频以外，包含的文字内容的形式可以为“上课科目为XX，A学生专注时长30分钟，专注力为集中，高于班上10％的同学，互动次数3次，微笑5次，特此提出表扬，愿继续努力”或是“上课科目为XX，B学生注意力较不集中，举手等手势互动频次较低，建议家长密切关注，及时调整孩子的学习习惯”等。In an example, in addition to pictures and videos, the visualized statistical data of learning status can contain text content in the form of "The subject of class is XX, the duration of concentration of A student is 30 minutes, and the concentration is concentrated, which is 10% higher than the class. % Of classmates interacted 3 times and smiled 5 times. I hereby give praise and are willing to continue to work hard" or "The subject of class is XX, B students have less concentration, and the frequency of gestures such as raising hands is lower. Parents are advised to pay close attention , Adjust the children’s study habits in time" and so on.
在本公开实施例中，通过获取至少一个目标对象所属的类别，从而生成至少一个类别的学习状态统计数据，和/或，将至少一个目标对象的统计结果进行可视化处理，生成至少一个目标对象的学习状态统计数据。通过上述过程，可以通过不同的数据统计方式，更为直观与全面地掌握目标对象的学习状态。In the embodiment of the present disclosure, by acquiring the category to which at least one target object belongs, the learning state statistical data of at least one category is generated, and/or the statistical result of the at least one target object is visualized to generate the statistics of the at least one target object. Statistics of learning status. Through the above process, the learning state of the target object can be grasped more intuitively and comprehensively through different statistical methods.
图2示出根据本公开实施例的视频处理装置的框图。如图所示，所述视频处理装置20可以包括：Fig. 2 shows a block diagram of a video processing device according to an embodiment of the present disclosure. As shown in the figure, the video processing device 20 may include:
视频获取模块21，用于获取视频，其中，视频中的至少部分视频帧包含目标对象；The video acquisition module 21 is configured to acquire a video, where at least part of the video frames in the video contain the target object;
检测模块22，用于根据视频，对目标对象在观看教学课程过程中的至少一类学习行为进行检测；The detection module 22 is used to detect at least one type of learning behavior of the target object in the process of watching the teaching course according to the video;
生成模块23，用于在检测到目标对象执行至少一类学习行为的情况下，根据至少部分包含至少一类学习行为的视频帧和/或目标对象执行至少一类学习行为的持续时间，生成学习状态信息。The generating module 23 is configured to generate learning based on at least part of the video frames containing at least one type of learning behavior and/or the duration of the target object performing at least one type of learning behavior when it is detected that the target object performs at least one type of learning behavior status information.
在一种可能的实现方式中，学习行为包括以下行为中的至少一类：执行至少一种目标手势、表现目标情绪、关注教学课程的展示区域、与其他对象产生至少一种互动行为、在视频中的至少部分视频帧中未出现、闭眼以及在教学课程的展示区域内的目光交流。In a possible implementation, the learning behavior includes at least one of the following behaviors: performing at least one target gesture, expressing the target emotion, paying attention to the display area of the teaching course, generating at least one interactive behavior with other objects, There is no eye contact, eyes closed, and eye contact in the display area of the teaching course in at least part of the video frames in.
在一种可能的实现方式中，检测模块用于：对视频进行目标对象检测，得到包含目标对象的视频帧；对包含目标对象的视频帧进行至少一类学习行为检测。In a possible implementation manner, the detection module is configured to: perform target object detection on the video to obtain a video frame containing the target object; and perform at least one type of learning behavior detection on the video frame containing the target object.
在一种可能的实现方式中，学习行为包括执行至少一种目标手势；检测模块进一步用于：对包含目标对象的视频帧进行至少一种目标手势的检测；在检测到包含至少一种目标手势的连续视频帧的数量超过第一阈值的情况下，将包含目标手势的视频帧中的至少一帧记录为手势开始帧；在手势开始帧以后的视频帧中，不包含目标手势的连续视频帧的数量超过第二阈值的情况下，将不包含目标手势的视频帧中的至少一帧记录为手势结束帧；根据手势开始帧与手势结束帧的数量，确定视频中目标对象执行至少一种目标手势的次数和/或时间。In a possible implementation manner, the learning behavior includes performing at least one target gesture; the detection module is further configured to: perform detection of at least one target gesture on the video frame containing the target object; When the number of continuous video frames exceeds the first threshold, record at least one of the video frames containing the target gesture as the gesture start frame; in the video frames after the gesture start frame, the continuous video frames that do not contain the target gesture When the number exceeds the second threshold, record at least one of the video frames that do not contain the target gesture as the gesture end frame; according to the number of gesture start frames and gesture end frames, determine that the target object in the video performs at least one target The number and/or time of gestures.
在一种可能的实现方式中，学习行为包括表现目标情绪；检测模块进一步用于：对包含目标对象的视频帧进行表情检测和/或微笑值检测；在检测到视频帧中目标对象展示至少一种第一目标表情或微笑值检测的结果超过目标微笑值情况下，将检测到的视频帧作为第一检测帧；在检测到连续的第一检测帧的数量超过第三阈值的情况下，确定目标对象产生目标情绪。In a possible implementation, the learning behavior includes expressing the target emotion; the detection module is further used to: perform expression detection and/or smile value detection on the video frame containing the target object; in the detected video frame, the target object displays at least one When the first target expression or smile value detection result exceeds the target smile value, the detected video frame is regarded as the first detection frame; when the number of consecutive first detection frames exceeds the third threshold, it is determined The target object produces the target emotion.
在一种可能的实现方式中，学习行为包括关注教学课程的展示区域；检测模块进一步用于：对包含目标对象的视频帧进行表情检测和人脸角度检测；在检测到视频帧中目标对象展示至少一种第二目标表情且人脸角度在目标人脸角度范围以内的情况下，将检测到的视频帧作为第二检测帧；在检测到连续的第二检测帧的数量超过第四阈值的情况下，确定目标对象关注教学课程的展示区域。In a possible implementation, the learning behavior includes paying attention to the display area of the teaching course; the detection module is further used to: perform expression detection and face angle detection on the video frame containing the target object; display the target object in the detected video frame In the case of at least one second target expression and the face angle is within the range of the target face angle, the detected video frame is used as the second detection frame; when the number of consecutive second detection frames exceeds the fourth threshold In this case, determine the target object to focus on the display area of the teaching course.
在一种可能的实现方式中，生成模块用于：获取视频中包含至少一类学习行为的视频帧，作为目标视频帧集合；对目标视频帧集合中的至少一个视频帧进行人脸质量检测，将人脸质量大于人脸质量阈值的视频帧作为目标视频帧；根据目标视频帧，生成学习状态信息。In a possible implementation, the generating module is used to: obtain video frames containing at least one type of learning behavior in the video as a target video frame set; perform face quality detection on at least one video frame in the target video frame set, The video frame whose face quality is greater than the face quality threshold is taken as the target video frame; according to the target video frame, the learning state information is generated.
在一种可能的实现方式中，生成模块进一步用于：将目标视频帧中的至少一帧作为学习状态信息；和/或，识别在至少一帧目标视频帧中目标对象所在区域，基于目标对象所在区域，生成学习状态信息。In a possible implementation manner, the generating module is further configured to: use at least one frame of the target video frame as the learning state information; and/or, identify the area where the target object is located in the at least one frame of the target video frame, based on the target object In the area, the learning status information is generated.
在一种可能的实现方式中，检测模块用于：对视频进行目标对象检测，得到包含目标对象的视频帧，并将视频中包含目标对象的视频帧以外的视频帧，作为未检测到目标对象的视频帧；在未检测到目标对象的视频帧的数量超过预设视频帧数量的情况下，检测到学习行为包括：在视频中的至少部分视频帧中未出现。In a possible implementation, the detection module is used to: perform target object detection on the video to obtain a video frame containing the target object, and use the video frame other than the video frame containing the target object as the undetected target object When the number of video frames in which the target object is not detected exceeds the preset number of video frames, the detected learning behavior includes: not appearing in at least part of the video frames in the video.
在一种可能的实现方式中，生成模块用于：在检测到目标对象执行至少一类学习行为的时间不小于时间阈值的情况下，记录至少一类学习行为的持续时间；将至少一类学习行为对应的持续时间，作为学习状态信息。In a possible implementation manner, the generating module is used to record the duration of at least one type of learning behavior when it is detected that the time for the target object to perform at least one type of learning behavior is not less than a time threshold; The duration corresponding to the behavior is used as the learning status information.
在一种可能的实现方式中，装置还用于：对视频中的至少部分视频帧中的背景区域进行渲染，其中，背景区域为视频帧中目标对象以外的区域。In a possible implementation manner, the device is further configured to: render a background area in at least part of the video frame in the video, where the background area is an area outside the target object in the video frame.
在一种可能的实现方式中，装置还用于：统计至少一个目标对象的学习状态信息，得到至少一个目标对象的统计结果；根据至少一个目标对象的统计结果，生成学习状态统计数据。In a possible implementation manner, the device is further configured to: collect statistics on the learning state information of at least one target object to obtain a statistical result of at least one target object; and generate statistical data of the learning state according to the statistical result of at least one target object.
在一种可能的实现方式中，装置还用于：根据至少一个目标对象所属的类别，获取至少一个类别包含的目标对象的统计结果，生成至少一个类别的学习状态统计数据，其中，目标对象所属的类别包括目标对象参与的课程、目标对象注册的机构以及目标对象使用的设备中的至少一种；和/或，将至少一个目标对象的统计结果进行可视化处理，生成至少一个目标对象的学习状态统计数据。In a possible implementation manner, the device is further configured to: obtain statistical results of the target objects contained in the at least one category according to the category to which the at least one target object belongs, and generate statistical data of the learning state of at least one category, wherein the target object belongs to The category includes at least one of the courses the target object participates in, the institution registered by the target object, and the equipment used by the target object; and/or visualize the statistical results of at least one target object to generate the learning status of at least one target object Statistical data.
在不违背逻辑的情况下，本申请不同实施例之间可以相互结合，不同实施例描述有所侧重，未侧重描述的部分可参见其他实施例的记载。Without violating logic, different embodiments of the present application can be combined with each other, and the description of different embodiments is emphasized, and the parts that are not described may be referred to the records of other embodiments.
在本公开的一些实施例中，本公开实施例提供的装置具有的功能或包含的模块可以用于执行上文方法实施例描述的方法，其具体实现和技术效果可参照上文方法实施例的描述，为了简洁，这里不再赘述。In some embodiments of the present disclosure, the functions or modules contained in the device provided in the embodiments of the present disclosure can be used to execute the methods described in the above method embodiments. For specific implementation and technical effects, please refer to the above method embodiments. Description, for the sake of brevity, I will not repeat it here.
应用场景示例Application scenario example
学生学习的方式通常是老师授课，学生听课，课堂缺少互动和趣味性，学生不容易提起兴趣听课，不能通过学生的实时表现对学生形成正向激励。同时，机构或者老师也无法掌握学生的听课状态，家长也无法了解孩子在学校的表现，尤其是受疫情影响，学生在线上课的时间非常多，然而，学生是否真正上课以及是否在认真听课、课堂互动表现如何，都无法量化评估。因此，如何有效地把握学生的学习状态，成为目前一个亟待解决的问题。The way students learn is usually the teacher teaches, the students listen to the lesson, the classroom lacks interaction and interest, the students are not easy to be interested in the lesson, and the real-time performance of the students cannot form a positive motivation for the students. At the same time, institutions or teachers cannot grasp the status of students’ attendance, and parents cannot understand their children’s performance at school. Especially affected by the epidemic, students spend a lot of time in online classes. However, whether students are actually attending classes and whether they are attending classes carefully. The performance of the interaction cannot be quantitatively evaluated. Therefore, how to effectively grasp the learning status of students has become an urgent problem to be solved at present.
本公开应用示例提出了一套学习系统，该系统可以通过上述公开实施例中提出的视频处理方法，来有效地掌握学生的学习状态。The application example of the present disclosure proposes a set of learning system, which can effectively grasp the learning state of students through the video processing method proposed in the above-mentioned disclosed embodiment.
图3示出根据本公开一应用示例的示意图。如图所示，在一个示例中，学习系统可以由用户端、教育软件服务化(SaaS，Software-as-a-Service)后台以及互动课堂后台等三部分所构成。其中，学生通过用户端观看教学课程，用户端可以包含两部分，分别是用于学习的硬件设备(比如图中安装了Windows系统或是IOS系统以及SDK的客户端)，以及学生登入在线课堂的应用程序(即图中的用户APP)。教育SaaS后台可以是学生所在的教育机构的服务器所搭建的平台，互动课堂后台可以是汇总不同教育机构的数据并进行数据维护的服务器所搭建的平台，无论是教育SaaS后台还是互动课堂后台，均可以通过API接口，与用户端之间进行数据交互。从而实现上述各公开实施例中所提到的学习状态信息生成以及学习状态统计数据的生成。Fig. 3 shows a schematic diagram of an application example according to the present disclosure. As shown in the figure, in an example, the learning system can be composed of three parts: the user end, the educational software service (SaaS, Software-as-a-Service) backend, and the interactive classroom backend. Among them, students watch the teaching courses through the client. The client can include two parts: hardware devices for learning (such as the client with Windows system or IOS system and SDK installed in the picture), and the student's login to the online classroom. Application (ie the user APP in the figure). The education SaaS backend can be a platform built by the server of the educational institution where the student is located, and the interactive classroom backend can be a platform built by a server that aggregates data from different educational institutions and performs data maintenance, whether it is an education SaaS backend or an interactive classroom backend. Data can be exchanged with the client through the API interface. Thereby, the generation of learning state information and the generation of learning state statistical data mentioned in the above disclosed embodiments are realized.
在本公开应用示例中，学习状态信息的生成过程可以包括：In the application example of the present disclosure, the process of generating learning state information may include:
用户端通过采集学生观看教学课程过程的视频，并对采集的视频进行处理，从而获取每个学生的学习状态信息，教育SaaS后台以及互动课堂后台通过API接口，调用不同用户端中生成的学习状态信息，并对这些学习状态信息通过上述公开实施例中提到的任意方式进行统计处理，生成学习状态统计数据。The user terminal obtains the learning status information of each student by collecting the videos of the students watching the teaching course process and processing the collected videos. The education SaaS background and the interactive classroom background call the learning status generated in different users through the API interface. Information, and perform statistical processing on the learning state information in any manner mentioned in the above-mentioned disclosed embodiment to generate learning state statistical data.
在一个示例中，用户端对采集的视频进行处理，获取每个学生的学习状态信息的过程可以包括：In an example, the user terminal processes the collected video, and the process of obtaining the learning status information of each student may include:
A.获取学生上课的精彩时刻(即上述公开实施例中提到的积极的学习行为)。A. Get the exciting moments of the students in class (that is, the positive learning behavior mentioned in the above-mentioned disclosed embodiment).
在一个示例中，可以通过定义一定的规则，制作学生的精彩视频集锦，可以将学生的表现剪辑成一小段视频或者是一些精彩图片并提供给家长，这样家长可以及时评估学生的上课表现，如果效果好，可能会鼓励学生继续参加相关课程。In an example, you can define certain rules to create a collection of exciting videos of students. You can edit the performance of students into a short video or some exciting pictures and provide them to parents, so that parents can evaluate students’ performance in class in time. Well, students may be encouraged to continue participating in related courses.
在一个示例中，获取学生的精彩时刻可以在学生签到成功后进行，后去的精彩时刻的视频或图片会上传后台或云端，同时，还可以选择学生是否实时可见上传的精彩时刻的内容。在一个示例中，精彩时刻定义规则可以包括：产生至少一种目标手势，目标手势可以包括举手、点赞、手势OK以及手势Yeah等，在一段时间范围内如果检测到学生执行以上的手势，则可以对包含有手势的视频进行图片或视频帧抽取。表现开心的目标情绪，在一段时间范围内如果检测到学生的表情是高兴，且微笑值达到某一目标微笑值(比如99分)，则可以有高兴标签的视频帧或是达到目标微笑值的视频帧进行图片或视频帧抽取。关注教学课程的展示区域，在一段时间范围内如果学生人脸朝向一直较正，即headpose在某个阈值范围内，则可以对这段时间范围内的视频进行图片或视频帧抽取。In an example, the student's wonderful moments can be obtained after the student signs in successfully, and the videos or pictures of the next wonderful moments will be uploaded to the background or the cloud. At the same time, it is also possible to choose whether the students can see the uploaded wonderful moments in real time. In one example, the highlight definition rule may include: generating at least one target gesture. The target gesture may include raising hand, like, gesture OK, gesture Yeah, etc. If a student is detected to perform the above gesture within a period of time, Then you can extract pictures or video frames from videos that contain gestures. Express the happy target emotion. If it is detected that the student’s expression is happy within a period of time, and the smile value reaches a certain target smile value (such as 99 points), there can be a video frame with a happy label or a target smile value. The video frame performs picture or video frame extraction. Pay attention to the display area of the teaching course. If the student's face orientation is always correct within a period of time, that is, the headpose is within a certain threshold range, then pictures or video frames can be extracted from the video within this period of time.
B.对学生的学习情况进行学情检测(针对上述公开实施例中提到的消极的学习行为)。B. Perform a learning situation test on the learning situation of the students (for the negative learning behavior mentioned in the above disclosed embodiment).
在一个示例中，可以将学生可能不在画面中，或者有不专注的情况，通过学情检测，将数据实时推送给家长，便于家长第一时间关注孩子，及时纠正孩子的不良学习习惯，起到辅助监督作用。In one example, the student may not be on the screen or may be unfocused, and the data can be pushed to the parents in real time through the learning situation detection, so that the parents can pay attention to the children in the first time, and correct the children’s bad learning habits in time. Auxiliary supervision.
在一个示例中，对学生进行学情检测的过程可以在学生签到成功后进行，如镜头前多长时间范围内无人出现、未观看屏幕、闭眼等，则判断该人专注度较低，在这种情况下，可以统计学生出现上述学习行为的时长，并将其作为学情检测的结果，得到相应的学习状态数据。具体的学情检测配置规 则可以参考上述各公开实施例，在此不再赘述。In one example, the process of checking the student's academic status can be carried out after the student signs in successfully. For example, for how long in front of the camera, no one appears in front of the camera, does not watch the screen, closes eyes, etc., it is judged that the person has a low degree of concentration. In this case, it is possible to count the length of time during which the student has the above-mentioned learning behavior, and use it as the result of the academic condition detection to obtain the corresponding learning state data. The specific academic condition detection configuration rules can refer to the above disclosed embodiments, which will not be repeated here.
通过上述各公开示例，可以得到包含有精彩时刻以及学情检测的学习状态信息，进一步地，教育SaaS后台以及互动课堂后台通过API接口，调用不同用户端中生成的学习状态信息，来生成学习状态统计数据的过程可以包括：Through the above public examples, learning status information including exciting moments and learning situation detection can be obtained. Further, the education SaaS backend and interactive classroom backend use API interfaces to call the learning status information generated in different client terminals to generate learning status. The process of statistical data can include:
C.报表生成(即上述公开实施例中的生成至少一个类别的学习状态统计数据)。C. Report generation (that is, the generation of statistical data of learning status in at least one category in the above disclosed embodiment).
在一个示例中，后台或云端API可以分设备、课程、机构等不同维度查看学生的签到信息以及学习状态信息，主要数据指标可以包括：签到时间、签到次数、比中人脸库(即上述公开实施例中的目标对象和预设数据库中的人脸匹配的情况)、签到设备、签到课程、专注时长以及微笑时长等。In one example, the backend or cloud API can view student sign-in information and learning status information in different dimensions such as device, course, institution, etc. The main data indicators can include: sign-in time, sign-in times, and face database (that is, the above-mentioned public The target object in the embodiment matches the face in the preset database), sign-in equipment, sign-in course, focus time, smile time, etc.
D.分析报告(即上述公开实施例中的可视化处理生成至少一个目标对象的学习状态统计数据)。D. Analysis report (that is, the visualization process in the above disclosed embodiment generates statistics on the learning status of at least one target object).
在一个示例中，教育SaaS后台或互动课堂后台可以将学生在在线课堂的表现情况，统一整理成一个完整的学情分析报告。报告通过可视化的图形界面说明学生上课的情况，进一步地，后台还可以选择较好的情况推送给家长或老师，从而可以用于机构老师分析学生情况，逐步辅助孩子改善学习行为。In one example, the education SaaS backend or the interactive classroom backend can unify the students' performance in the online classroom into a complete academic analysis report. The report explains the student’s class status through a visual graphical interface. Furthermore, the background can also select a better situation and push it to parents or teachers, so that it can be used by institutional teachers to analyze the student’s situation and gradually assist children in improving their learning behavior.
除上述过程以外，学习系统还可以在学生通过用户端进行学习的过程中，对学生的学习视频进行背景分割处理。在一个示例中，用户端可以针对于学生没有适合直播的位置背景或者出于隐私保护不愿意显示背景画面的情况，提供背景分割功能。在一个示例中，用户端的SDK可以支持若干不同的背景模版，比如可以预设置若干通用模版，在一个示例中，学生也可通过用户端从互动课堂后台调用定制模版。在一个示例中，SDK可以提供背景模版预览接口给用户端的APP，便于学生通过APP预览可以调用的定制模板；学生在上课过程中，也可以通过用户端中APP上背景分割的贴纸，用于对直播背景进行渲染，在一个示例中，如果学生不满意贴纸，也可以手动触发关闭。用户端的APP可以将学生使用贴纸的数据上报相应后台(教育SaaS后台或互动课堂后台)，相应后台可以分析学生使用了哪些背景贴纸以及使用量等信息，作为额外的学习状态信息等。In addition to the above process, the learning system can also perform background segmentation processing on the student's learning video when the student is learning through the user terminal. In one example, the user terminal may provide a background segmentation function for situations where the student does not have a location background suitable for live broadcast or is unwilling to display a background image for privacy protection. In an example, the SDK on the user side can support several different background templates. For example, several general templates can be preset. In one example, students can also call customized templates from the interactive classroom backend through the user side. In one example, the SDK can provide a background template preview interface to the app on the user side, so that students can preview the customized templates that can be called through the app; students can also use the background segmentation stickers on the app on the user side to compare The live broadcast background is rendered. In one example, if the student is not satisfied with the sticker, it can also be manually triggered to close. The APP on the user side can report the data of students using stickers to the corresponding back-end (education SaaS back-end or interactive classroom back-end), and the corresponding back-end can analyze which background stickers are used by students and information such as usage amount as additional learning status information.
本公开应用示例中提出的学习系统，除了可以应用于在线课堂外，还可以扩展应用于其他相关领域，比如在线会议等。The learning system proposed in the application examples of the present disclosure can not only be applied to online classrooms, but also be extended to other related fields, such as online meetings.
可以理解，本公开提及的上述各个方法实施例，在不违背原理逻辑的情况下，均可以彼此相互结合形成结合后的实施例，限于篇幅，本公开不再赘述。It can be understood that the various method embodiments mentioned in the present disclosure can be combined with each other to form a combined embodiment without violating the principle and logic. The length is limited, and the details of this disclosure will not be repeated.
本领域技术人员可以理解，在具体实施方式的上述方法中，各步骤的撰写顺序并不意味着严格的执行顺序而对实施过程构成任何限定，各步骤的具体执行顺序应当以其功能和可能的内在逻辑确定。Those skilled in the art can understand that in the above-mentioned methods of the specific implementation, the writing order of the steps does not mean a strict execution order but constitutes any limitation on the implementation process. The specific execution order of each step should be based on its function and possibility. The inner logic is determined.
本公开实施例还提出一种计算机可读存储介质，其上存储有计算机程序指令，所述计算机程序指令被处理器执行时实现上述方法。计算机可读存储介质可以是易失性计算机可读存储介质或非易失性计算机可读存储介质。The embodiments of the present disclosure also provide a computer-readable storage medium on which computer program instructions are stored, and the computer program instructions implement the above-mentioned method when executed by a processor. The computer-readable storage medium may be a volatile computer-readable storage medium or a non-volatile computer-readable storage medium.
本公开实施例还提出一种电子设备，包括：处理器；用于存储处理器可执行指令的存储器；其中，所述处理器被配置为上述方法。An embodiment of the present disclosure also provides an electronic device, including: a processor; a memory for storing executable instructions of the processor; wherein the processor is configured as the above-mentioned method.
本公开实施例还提出一种计算机程序，包括计算机可读代码，当所述计算机可读代码在电子设备中运行时，所述电子设备中的处理器执行用于实现上述方法。The embodiment of the present disclosure also provides a computer program, including computer readable code, when the computer readable code is executed in an electronic device, the processor in the electronic device is executed to implement the above method.
在实际应用中，上述存储器可以是易失性存储器(volatile memory)，例如RAM；或者非易失性存储器(non-volatile memory)，例如ROM，快闪存储器(flash memory)，硬盘(Hard Disk Drive，HDD)或固态硬盘(Solid-State Drive，SSD)；或者上述种类的存储器的组合，并向处理器提供指令和数据。In practical applications, the above-mentioned memory may be a volatile memory (volatile memory), such as RAM; or a non-volatile memory (non-volatile memory), such as ROM, flash memory, hard disk drive (Hard Disk Drive) , HDD) or solid-state drive (Solid-State Drive, SSD); or a combination of the above types of memory, and provide instructions and data to the processor.
上述处理器可以为ASIC、DSP、DSPD、PLD、FPGA、CPU、控制器、微控制器、微处理器中的至少一种。可以理解地，对于不同的设备，用于实现上述处理器功能的电子器件还可以为其它，本公开实施例不作具体限定。The foregoing processor may be at least one of ASIC, DSP, DSPD, PLD, FPGA, CPU, controller, microcontroller, and microprocessor. It is understandable that, for different devices, the electronic device used to implement the above-mentioned processor function may also be other, and the embodiment of the present disclosure does not specifically limit it.
电子设备可以被提供为终端、服务器或其它形态的设备。The electronic device can be provided as a terminal, server or other form of device.
基于前述实施例相同的技术构思，本公开实施例还提供了一种计算机程序，该计算机程序被处理器执行时实现上述方法。Based on the same technical concept as the foregoing embodiment, the embodiment of the present disclosure also provides a computer program, which implements the foregoing method when the computer program is executed by a processor.
图4是根据本公开实施例的一种电子设备800的框图。例如，电子设备800可以是移动电话，计算机，数字广播终端，消息收发设备，游戏控制台，平板设备，医疗设备，健身设备，个人数字助理等终端。FIG. 4 is a block diagram of an electronic device 800 according to an embodiment of the present disclosure. For example, the electronic device 800 may be a mobile phone, a computer, a digital broadcasting terminal, a messaging device, a game console, a tablet device, a medical device, a fitness device, a personal digital assistant, and other terminals.
参照图4，电子设备800可以包括以下一个或多个组件：处理组件802，存储器804，电源组件806，多媒体组件808，音频组件810，输入/输出(I/O)的接口812，传感器组件814，以及通信组件816。4, the electronic device 800 may include one or more of the following components: a processing component 802, a memory 804, a power component 806, a multimedia component 808, an audio component 810, an input/output (I/O) interface 812, and a sensor component 814 , And communication component 816.
处理组件802通常控制电子设备800的整体操作，诸如与显示，电话呼叫，数据通信，相机操作和记录操作相关联的操作。处理组件802可以包括一个或多个处理器820来执行指令，以完成上述的方法的全部或部分步骤。此外，处理组件802可以包括一个或多个模块，便于处理组件802和其他组件之间的交互。例如，处理组件802可以包括多媒体模块，以方便多媒体组件808和处理组件802之间的交互。The processing component 802 generally controls the overall operations of the electronic device 800, such as operations associated with display, telephone calls, data communications, camera operations, and recording operations. The processing component 802 may include one or more processors 820 to execute instructions to complete all or part of the steps of the foregoing method. In addition, the processing component 802 may include one or more modules to facilitate the interaction between the processing component 802 and other components. For example, the processing component 802 may include a multimedia module to facilitate the interaction between the multimedia component 808 and the processing component 802.
存储器804被配置为存储各种类型的数据以支持在电子设备800的操作。这些数据的示例包括用于在电子设备800上操作的任何应用程序或方法的指令，联系人数据，电话簿数据，消息，图片，视频等。存储器804可以由任何类型的易失性或非易失性存储设备或者它们的组合实现，如静态随机存取存储器(SRAM)，电可擦除可编程只读存储器(EEPROM)，可擦除可编程只读存储器(EPROM)，可编程只读存储器(PROM)，只读存储器(ROM)，磁存储器，快闪存储器，磁盘或光盘。The memory 804 is configured to store various types of data to support operations in the electronic device 800. Examples of these data include instructions for any application or method to operate on the electronic device 800, contact data, phone book data, messages, pictures, videos, etc. The memory 804 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable and Programmable Read Only Memory (EPROM), Programmable Read Only Memory (PROM), Read Only Memory (ROM), Magnetic Memory, Flash Memory, Magnetic Disk or Optical Disk.
电源组件806为电子设备800的各种组件提供电力。电源组件806可以包括电源管理系统，一个或多个电源，及其他与为电子设备800生成、管理和分配电力相关联的组件。The power supply component 806 provides power for various components of the electronic device 800. The power supply component 806 may include a power management system, one or more power supplies, and other components associated with the generation, management, and distribution of power for the electronic device 800.
多媒体组件808包括在所述电子设备800和用户之间的提供一个输出接口的屏幕。在一些实施例中，屏幕可以包括液晶显示器(LCD)和触摸面板(TP)。如果屏幕包括触摸面板，屏幕可以被实现为触摸屏，以接收来自用户的输入信号。触摸面板包括一个或多个触摸传感器以感测触摸、滑动和触摸面板上的手势。所述触摸传感器可以不仅感测触摸或滑动动作的边界，而且还检测与所述触摸或滑动操作相关的持续时间和压力。在一些实施例中，多媒体组件808包括一个前置摄像头和/或后置摄像头。当电子设备800处于操作模式，如拍摄模式或视频模式时，前置摄像头和/或后置摄像头可以接收外部的多媒体数据。每个前置摄像头和后置摄像头可以是一个固定的光学透镜系统或具有焦距和光学变焦能力。The multimedia component 808 includes a screen that provides an output interface between the electronic device 800 and the user. In some embodiments, the screen may include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen may be implemented as a touch screen to receive input signals from the user. The touch panel includes one or more touch sensors to sense touch, sliding, and gestures on the touch panel. The touch sensor may not only sense the boundary of a touch or slide action, but also detect the duration and pressure related to the touch or slide operation. In some embodiments, the multimedia component 808 includes a front camera and/or a rear camera. When the electronic device 800 is in an operation mode, such as a shooting mode or a video mode, the front camera and/or the rear camera can receive external multimedia data. Each front camera and rear camera can be a fixed optical lens system or have focal length and optical zoom capabilities.
音频组件810被配置为输出和/或输入音频信号。例如，音频组件810包括一个麦克风(MIC)，当电子设备800处于操作模式，如呼叫模式、记录模式和语音识别模式时，麦克风被配置为接收外部音频信号。所接收的音频信号可以被进一步存储在存储器804或经由通信组件816发送。在一些实施例中，音频组件810还包括一个扬声器，用于输出音频信号。The audio component 810 is configured to output and/or input audio signals. For example, the audio component 810 includes a microphone (MIC), and when the electronic device 800 is in an operation mode, such as a call mode, a recording mode, and a voice recognition mode, the microphone is configured to receive an external audio signal. The received audio signal may be further stored in the memory 804 or transmitted via the communication component 816. In some embodiments, the audio component 810 further includes a speaker for outputting audio signals.
I/O接口812为处理组件802和外围接口模块之间提供接口，上述外围接口模块可以是键盘，点击轮，按钮等。这些按钮可包括但不限于：主页按钮、音量按钮、启动按钮和锁定按钮。The I/O interface 812 provides an interface between the processing component 802 and a peripheral interface module. The above-mentioned peripheral interface module may be a keyboard, a click wheel, a button, and the like. These buttons may include, but are not limited to: home button, volume button, start button, and lock button.
传感器组件814包括一个或多个传感器，用于为电子设备800提供各个方面的状态评估。例如，传感器组件814可以检测到电子设备800的打开/关闭状态，组件的相对定位，例如所述组件为电子设备800的显示器和小键盘，传感器组件814还可以检测电子设备800或电子设备800一个组件的位置改变，用户与电子设备800接触的存在或不存在，电子设备800方位或加速/减速和电子设备800的温度变化。传感器组件814可以包括接近传感器，被配置用来在没有任何的物理接触时检测附近物体的存在。传感器组件814还可以包括光传感器，如CMOS或CCD图像传感器，用于在成像应用中使用。在一些实施例中，该传感器组件814还可以包括加速度传感器，陀螺仪传感器，磁传感器，压力传感器或温度传感器。The sensor component 814 includes one or more sensors for providing the electronic device 800 with various aspects of state evaluation. For example, the sensor component 814 can detect the on/off status of the electronic device 800 and the relative positioning of the components. For example, the component is the display and the keypad of the electronic device 800. The sensor component 814 can also detect the electronic device 800 or the electronic device 800. The position of the component changes, the presence or absence of contact between the user and the electronic device 800, the orientation or acceleration/deceleration of the electronic device 800, and the temperature change of the electronic device 800. The sensor component 814 may include a proximity sensor configured to detect the presence of nearby objects when there is no physical contact. The sensor component 814 may also include a light sensor, such as a CMOS or CCD image sensor, for use in imaging applications. In some embodiments, the sensor component 814 may also include an acceleration sensor, a gyroscope sensor, a magnetic sensor, a pressure sensor, or a temperature sensor.
通信组件816被配置为便于电子设备800和其他设备之间有线或无线方式的通信。电子设备800可以接入基于通信标准的无线网络，如WiFi，2G，3G，4G或5G，或它们的组合。在一个示例性实施例中，通信组件816经由广播信道接收来自外部广播管理系统的广播信号或广播相关人员信息。在一个示例性实施例中，所述通信组件816还包括近场通信(NFC)模块，以促进短程通信。例如，在NFC模块可基于射频识别(RFID)技术，红外数据协会(IrDA)技术，超宽带(UWB)技术，蓝牙(BT) 技术和其他技术来实现。The communication component 816 is configured to facilitate wired or wireless communication between the electronic device 800 and other devices. The electronic device 800 can access a wireless network based on a communication standard, such as WiFi, 2G, 3G, 4G, or 5G, or a combination thereof. In an exemplary embodiment, the communication component 816 receives a broadcast signal or broadcast related personnel information from an external broadcast management system via a broadcast channel. In an exemplary embodiment, the communication component 816 further includes a near field communication (NFC) module to facilitate short-range communication. For example, the NFC module can be implemented based on radio frequency identification (RFID) technology, infrared data association (IrDA) technology, ultra-wideband (UWB) technology, Bluetooth (BT) technology and other technologies.
在示例性实施例中，电子设备800可以被一个或多个应用专用集成电路(ASIC)、数字信号处理器(DSP)、数字信号处理设备(DSPD)、可编程逻辑器件(PLD)、现场可编程门阵列(FPGA)、控制器、微控制器、微处理器或其他电子元件实现，用于执行上述方法。In an exemplary embodiment, the electronic device 800 may be implemented by one or more application-specific integrated circuits (ASIC), digital signal processors (DSP), digital signal processing devices (DSPD), programmable logic devices (PLD), field-available A programmable gate array (FPGA), controller, microcontroller, microprocessor, or other electronic components are implemented to implement the above methods.
在示例性实施例中，还提供了一种非易失性计算机可读存储介质，例如包括计算机程序指令的存储器804，上述计算机程序指令可由电子设备800的处理器820执行以完成上述方法。In an exemplary embodiment, there is also provided a non-volatile computer-readable storage medium, such as the memory 804 including computer program instructions, which can be executed by the processor 820 of the electronic device 800 to complete the foregoing method.
图5是根据本公开实施例的一种电子设备1900的框图。例如，电子设备1900可以被提供为一服务器。参照图5，电子设备1900包括处理组件1922，其进一步包括一个或多个处理器，以及由存储器1932所代表的存储器资源，用于存储可由处理组件1922的执行的指令，例如应用程序。存储器1932中存储的应用程序可以包括一个或一个以上的每一个对应于一组指令的模块。此外，处理组件1922被配置为执行指令，以执行上述方法。FIG. 5 is a block diagram of an electronic device 1900 according to an embodiment of the present disclosure. For example, the electronic device 1900 may be provided as a server. 5, the electronic device 1900 includes a processing component 1922, which further includes one or more processors, and a memory resource represented by a memory 1932, for storing instructions executable by the processing component 1922, such as application programs. The application program stored in the memory 1932 may include one or more modules each corresponding to a set of instructions. In addition, the processing component 1922 is configured to execute instructions to perform the above-described methods.
电子设备1900还可以包括一个电源组件1926被配置为执行电子设备1900的电源管理，一个有线或无线网络接口1950被配置为将电子设备1900连接到网络，和一个输入输出(I/O)接口1958。电子设备1900可以操作基于存储在存储器1932的操作系统，例如Windows ServerTM，Mac OS XTM，UnixTM,LinuxTM，FreeBSDTM或类似。The electronic device 1900 may also include a power supply component 1926 configured to perform power management of the electronic device 1900, a wired or wireless network interface 1950 configured to connect the electronic device 1900 to the network, and an input output (I/O) interface 1958 . The electronic device 1900 can operate based on an operating system stored in the memory 1932, such as Windows ServerTM, Mac OS XTM, UnixTM, LinuxTM, FreeBSDTM or the like.
在示例性实施例中，还提供了一种非易失性计算机可读存储介质，例如包括计算机程序指令的存储器1932，上述计算机程序指令可由电子设备1900的处理组件1922执行以完成上述方法。In an exemplary embodiment, a non-volatile computer-readable storage medium is also provided, such as the memory 1932 including computer program instructions, which can be executed by the processing component 1922 of the electronic device 1900 to complete the foregoing method.
本公开可以是系统、方法和/或计算机程序产品。计算机程序产品可以包括计算机可读存储介质，其上载有用于使处理器实现本公开的各个方面的计算机可读程序指令。The present disclosure may be a system, method and/or computer program product. The computer program product may include a computer-readable storage medium loaded with computer-readable program instructions for enabling a processor to implement various aspects of the present disclosure.
计算机可读存储介质可以是可以保持和存储由指令执行设备使用的指令的有形设备。计算机可读存储介质例如可以是――但不限于――电存储设备、磁存储设备、光存储设备、电磁存储设备、半导体存储设备或者上述的任意合适的组合。计算机可读存储介质的更具体的例子(非穷举的列表)包括：便携式计算机盘、硬盘、随机存取存储器(RAM)、只读存储器(ROM)、可擦式可编程只读存储器(EPROM或闪存)、静态随机存取存储器(SRAM)、便携式压缩盘只读存储器(CD-ROM)、数字多功能盘(DVD)、记忆棒、软盘、机械编码设备、例如其上存储有指令的打孔卡或凹槽内凸起结构、以及上述的任意合适的组合。这里所使用的计算机可读存储介质不被解释为瞬时信号本身，诸如无线电波或者其他自由传播的电磁波、通过波导或其他传输媒介传播的电磁波(例如，通过光纤电缆的光脉冲)、或者通过电线传输的电信号。The computer-readable storage medium may be a tangible device that can hold and store instructions used by the instruction execution device. The computer-readable storage medium may be, for example, but not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing. More specific examples (non-exhaustive list) of computer-readable storage media include: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM) Or flash memory), static random access memory (SRAM), portable compact disk read-only memory (CD-ROM), digital versatile disk (DVD), memory stick, floppy disk, mechanical encoding device, such as a printer with instructions stored thereon The protruding structure in the hole card or the groove, and any suitable combination of the above. The computer-readable storage medium used here is not interpreted as the instantaneous signal itself, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (for example, light pulses through fiber optic cables), or through wires Transmission of electrical signals.
这里所描述的计算机可读程序指令可以从计算机可读存储介质下载到各个计算/处理设备，或者通过网络、例如因特网、局域网、广域网和/或无线网下载到外部计算机或外部存储设备。网络可以包括铜传输电缆、光纤传输、无线传输、路由器、防火墙、交换机、网关计算机和/或边缘服务器。每个计算/处理设备中的网络适配卡或者网络接口从网络接收计算机可读程序指令，并转发该计算机可读程序指令，以供存储在各个计算/处理设备中的计算机可读存储介质中。The computer-readable program instructions described herein can be downloaded from a computer-readable storage medium to various computing/processing devices, or downloaded to an external computer or external storage device via a network, such as the Internet, a local area network, a wide area network, and/or a wireless network. The network may include copper transmission cables, optical fiber transmission, wireless transmission, routers, firewalls, switches, gateway computers, and/or edge servers. The network adapter card or network interface in each computing/processing device receives computer-readable program instructions from the network, and forwards the computer-readable program instructions for storage in the computer-readable storage medium in each computing/processing device .
用于执行本公开操作的计算机程序指令可以是汇编指令、指令集架构(ISA)指令、机器指令、机器相关指令、微代码、固件指令、状态设置数据、或者以一种或多种编程语言的任意组合编写的源代码或目标代码，所述编程语言包括面向对象的编程语言—诸如Smalltalk、C++等，以及常规的过程式编程语言—诸如“C”语言或类似的编程语言。计算机可读程序指令可以完全地在用户计算机上执行、部分地在用户计算机上执行、作为一个独立的软件包执行、部分在用户计算机上部分在远程计算机上执行、或者完全在远程计算机或服务器上执行。在涉及远程计算机的情形中，远程计算机可以通过任意种类的网络—包括局域网(LAN)或广域网(WAN)—连接到用户计算机，或者，可以连接到外部计算机(例如利用因特网服务提供商来通过因特网连接)。在一些实施例中，通过利用计算机可读程序指令的状态人员信息来个性化定制电子电路，例如可编程逻辑电路、现场可编程门阵列(FPGA)或可编程逻辑阵列(PLA)，该电子电路可以执行计算机可读程序指令，从而实现本公开的各个方面。The computer program instructions used to perform the operations of the present disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-related instructions, microcode, firmware instructions, state setting data, or in one or more programming languages. Source code or object code written in any combination, the programming language includes object-oriented programming languages such as Smalltalk, C++, etc., and conventional procedural programming languages such as "C" language or similar programming languages. Computer-readable program instructions can be executed entirely on the user's computer, partly on the user's computer, executed as a stand-alone software package, partly on the user's computer and partly executed on a remote computer, or entirely on the remote computer or server implement. In the case of a remote computer, the remote computer can be connected to the user's computer through any kind of network-including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computer (for example, using an Internet service provider to connect to the user's computer) connect). In some embodiments, an electronic circuit, such as a programmable logic circuit, a field programmable gate array (FPGA), or a programmable logic array (PLA), is personalized by using status personnel information of computer-readable program instructions. The computer-readable program instructions can be executed to implement various aspects of the present disclosure.
这里参照根据本公开实施例的方法、装置(系统)和计算机程序产品的流程图和/或框图描述了本公开的各个方面。应当理解，流程图和/或框图的每个方框以及流程图和/或框图中各方框的组合， 都可以由计算机可读程序指令实现。Here, various aspects of the present disclosure are described with reference to flowcharts and/or block diagrams of methods, devices (systems) and computer program products according to embodiments of the present disclosure. It should be understood that each block of the flowcharts and/or block diagrams, and combinations of blocks in the flowcharts and/or block diagrams, can be implemented by computer-readable program instructions.
这些计算机可读程序指令可以提供给通用计算机、专用计算机或其它可编程数据处理装置的处理器，从而生产出一种机器，使得这些指令在通过计算机或其它可编程数据处理装置的处理器执行时，产生了实现流程图和/或框图中的一个或多个方框中规定的功能/动作的装置。也可以把这些计算机可读程序指令存储在计算机可读存储介质中，这些指令使得计算机、可编程数据处理装置和/或其他设备以特定方式工作，从而，存储有指令的计算机可读介质则包括一个制造品，其包括实现流程图和/或框图中的一个或多个方框中规定的功能/动作的各个方面的指令。These computer-readable program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, thereby producing a machine that makes these instructions when executed by the processor of the computer or other programmable data processing device , A device that implements the functions/actions specified in one or more blocks in the flowcharts and/or block diagrams is produced. It is also possible to store these computer-readable program instructions in a computer-readable storage medium. These instructions make computers, programmable data processing apparatuses, and/or other devices work in a specific manner. Thus, the computer-readable medium storing the instructions includes An article of manufacture, which includes instructions for implementing various aspects of the functions/actions specified in one or more blocks in the flowcharts and/or block diagrams.
也可以把计算机可读程序指令加载到计算机、其它可编程数据处理装置、或其它设备上，使得在计算机、其它可编程数据处理装置或其它设备上执行一系列操作步骤，以产生计算机实现的过程，从而使得在计算机、其它可编程数据处理装置、或其它设备上执行的指令实现流程图和/或框图中的一个或多个方框中规定的功能/动作。It is also possible to load computer-readable program instructions on a computer, other programmable data processing device, or other equipment, so that a series of operation steps are executed on the computer, other programmable data processing device, or other equipment to produce a computer-implemented process , So that the instructions executed on the computer, other programmable data processing apparatus, or other equipment realize the functions/actions specified in one or more blocks in the flowcharts and/or block diagrams.
附图中的流程图和框图显示了根据本公开的多个实施例的系统、方法和计算机程序产品的可能实现的体系架构、功能和操作。在这点上，流程图或框图中的每个方框可以代表一个模块、程序段或指令的一部分，所述模块、程序段或指令的一部分包含一个或多个用于实现规定的逻辑功能的可执行指令。在有些作为替换的实现中，方框中所标注的功能也可以以不同于附图中所标注的顺序发生。例如，两个连续的方框实际上可以基本并行地执行，它们有时也可以按相反的顺序执行，这依所涉及的功能而定。也要注意的是，框图和/或流程图中的每个方框、以及框图和/或流程图中的方框的组合，可以用执行规定的功能或动作的专用的基于硬件的系统来实现，或者可以用专用硬件与计算机指令的组合来实现。The flowcharts and block diagrams in the accompanying drawings show the possible implementation architecture, functions, and operations of the system, method, and computer program product according to multiple embodiments of the present disclosure. In this regard, each block in the flowchart or block diagram may represent a module, program segment, or part of an instruction, and the module, program segment, or part of an instruction contains one or more components for realizing the specified logical function. Executable instructions. In some alternative implementations, the functions marked in the block may also occur in a different order from the order marked in the drawings. For example, two consecutive blocks can actually be executed substantially in parallel, or they can sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and/or flowchart, and the combination of the blocks in the block diagram and/or flowchart, can be implemented by a dedicated hardware-based system that performs the specified functions or actions Or it can be realized by a combination of dedicated hardware and computer instructions.
以上已经描述了本公开的各实施例，上述说明是示例性的，并非穷尽性的，并且也不限于所披露的各实施例。在不偏离所说明的各实施例的范围和精神的情况下，对于本技术领域的普通技术人员来说许多修改和变更都是显而易见的。本文中所用术语的选择，旨在最好地解释各实施例的原理、实际应用或对市场中的技术改进，或者使本技术领域的其它普通技术人员能理解本文披露的各实施例。The embodiments of the present disclosure have been described above, and the above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Without departing from the scope and spirit of the illustrated embodiments, many modifications and changes are obvious to those of ordinary skill in the art. The choice of terms used herein is intended to best explain the principles, practical applications, or technical improvements in the market of the various embodiments, or to enable other ordinary skilled in the art to understand the various embodiments disclosed herein.

Claims (17)

  1. 一种视频处理方法，其特征在于，包括：A video processing method, characterized in that it comprises:
    获取视频，其中，所述视频中的至少部分视频帧包含目标对象；Acquiring a video, where at least part of the video frames in the video contains the target object;
    根据所述视频，对所述目标对象在观看教学课程过程中的至少一类学习行为进行检测；According to the video, detect at least one type of learning behavior of the target object in the process of watching the teaching course;
    在检测到所述目标对象执行至少一类学习行为的情况下，根据至少部分包含所述至少一类学习行为的视频帧和/或所述目标对象执行所述至少一类学习行为的持续时间，生成学习状态信息。In the case of detecting that the target object performs at least one type of learning behavior, according to at least a part of the video frame containing the at least one type of learning behavior and/or the duration of the target object performing the at least one type of learning behavior, Generate learning status information.
  2. 根据权利要求1所述的方法，其特征在于，所述学习行为包括以下行为中的至少一类：执行至少一种目标手势、表现目标情绪、关注所述教学课程的展示区域、与其他对象产生至少一种互动行为、在所述视频中的至少部分视频帧中未出现、闭眼以及在所述教学课程的展示区域内的目光交流。The method according to claim 1, wherein the learning behavior includes at least one of the following behaviors: performing at least one target gesture, expressing target emotions, paying attention to the display area of the teaching course, and producing with other objects. At least one interactive behavior, not appearing in at least part of the video frames in the video, eyes closed, and eye contact in the display area of the teaching course.
  3. 根据权利要求1或2所述的方法，其特征在于，所述根据所述视频，对所述目标对象的至少一类学习行为进行检测，包括：The method according to claim 1 or 2, wherein the detecting at least one type of learning behavior of the target object according to the video comprises:
    对所述视频进行目标对象检测，得到包含所述目标对象的视频帧；Performing target object detection on the video to obtain a video frame containing the target object;
    对包含所述目标对象的视频帧进行至少一类学习行为检测。At least one type of learning behavior detection is performed on the video frame containing the target object.
  4. 根据权利要求3所述的方法，其特征在于，所述学习行为包括执行至少一种目标手势；The method according to claim 3, wherein the learning behavior comprises performing at least one target gesture;
    所述对包含所述目标对象的视频帧进行至少一类学习行为检测，包括：The performing at least one type of learning behavior detection on the video frame containing the target object includes:
    对包含所述目标对象的视频帧进行至少一种目标手势的检测；Detecting at least one target gesture on the video frame containing the target object;
    在检测到包含至少一种所述目标手势的连续视频帧的数量超过第一阈值的情况下，将包含所述目标手势的视频帧中的至少一帧记录为手势开始帧；In a case where it is detected that the number of continuous video frames containing at least one of the target gestures exceeds the first threshold, at least one of the video frames containing the target gesture is recorded as a gesture start frame;
    在手势开始帧以后的视频帧中，不包含所述目标手势的连续视频帧的数量超过第二阈值的情况下，将不包含所述目标手势的视频帧中的至少一帧记录为手势结束帧；In the video frames after the gesture start frame, if the number of consecutive video frames that do not include the target gesture exceeds the second threshold, record at least one of the video frames that do not include the target gesture as the gesture end frame ；
    根据所述手势开始帧与所述手势结束帧的数量，确定所述视频中所述目标对象执行至少一种目标手势的次数和/或时间。According to the number of the gesture start frame and the gesture end frame, determine the number of times and/or time for the target object in the video to perform at least one target gesture.
  5. 根据权利要求3或4所述的方法，其特征在于，所述学习行为包括表现目标情绪；The method according to claim 3 or 4, wherein the learning behavior includes expressing a target emotion;
    所述对包含所述目标对象的视频帧进行至少一类学习行为检测，包括：The performing at least one type of learning behavior detection on the video frame containing the target object includes:
    对包含所述目标对象的视频帧进行表情检测和/或微笑值检测；Performing expression detection and/or smile value detection on the video frame containing the target object;
    在检测到视频帧中所述目标对象展示至少一种第一目标表情或微笑值检测的结果超过目标微笑值情况下，将检测到的视频帧作为第一检测帧；In a case where it is detected that the target object in the video frame shows at least one first target expression or a smile value detection result exceeds the target smile value, use the detected video frame as the first detection frame;
    在检测到连续的所述第一检测帧的数量超过第三阈值的情况下，确定所述目标对象产生所述目标情绪。In a case where it is detected that the number of consecutive first detection frames exceeds a third threshold, it is determined that the target object generates the target emotion.
  6. 根据权利要求3至5中任意一项所述的方法，其特征在于，所述学习行为包括关注所述教学课程的展示区域；The method according to any one of claims 3 to 5, wherein the learning behavior includes paying attention to the display area of the teaching course;
    所述对包含所述目标对象的视频帧进行至少一类学习行为检测，包括：The performing at least one type of learning behavior detection on the video frame containing the target object includes:
    对包含所述目标对象的视频帧进行表情检测和人脸角度检测；Performing expression detection and face angle detection on the video frame containing the target object;
    在检测到视频帧中所述目标对象展示至少一种第二目标表情且人脸角度在目标人脸角度范围以内的情况下，将检测到的视频帧作为第二检测帧；In a case where it is detected that the target object in the video frame shows at least one second target expression and the face angle is within the target face angle range, use the detected video frame as the second detection frame;
    在检测到连续的所述第二检测帧的数量超过第四阈值的情况下，确定所述目标对象关注所述教学课程的展示区域。In a case where it is detected that the number of consecutive second detection frames exceeds a fourth threshold, it is determined that the target object pays attention to the display area of the teaching course.
  7. 根据权利要求1至6中任意一项所述的方法，其特征在于，所述根据至少部分包含所述至少一类学习行为的视频帧，生成学习状态信息，包括：The method according to any one of claims 1 to 6, wherein the generating learning state information according to a video frame at least partially containing the at least one type of learning behavior comprises:
    获取所述视频中包含至少一类学习行为的视频帧，作为目标视频帧集合；Acquiring video frames containing at least one type of learning behavior in the video as a target video frame set;
    对所述目标视频帧集合中的至少一个视频帧进行人脸质量检测，将人脸质量大于人脸质量阈值的视频帧作为目标视频帧；Perform face quality detection on at least one video frame in the target video frame set, and use a video frame with a face quality greater than a face quality threshold as a target video frame;
    根据所述目标视频帧，生成所述学习状态信息。According to the target video frame, the learning state information is generated.
  8. 根据权利要求7所述的方法，其特征在于，所述根据所述目标视频帧，生成所述学习状态信息，包括：The method according to claim 7, wherein said generating said learning state information according to said target video frame comprises:
    将所述目标视频帧中的至少一帧作为学习状态信息；和/或，Use at least one of the target video frames as learning state information; and/or,
    识别在至少一帧所述目标视频帧中所述目标对象所在区域，基于所述目标对象所在区域，生成所述学习状态信息。Identify the area where the target object is located in at least one frame of the target video frame, and generate the learning state information based on the area where the target object is located.
  9. 根据权利要求1或2所述的方法，其特征在于，所述根据所述视频，对所述目标对象的至少一类学习行为进行检测，包括：The method according to claim 1 or 2, wherein the detecting at least one type of learning behavior of the target object according to the video comprises:
    对所述视频进行目标对象检测，得到包含所述目标对象的视频帧，并将所述视频中包含所述目标对象的视频帧以外的视频帧，作为未检测到目标对象的视频帧；Performing target object detection on the video to obtain a video frame containing the target object, and using a video frame other than the video frame containing the target object in the video as a video frame in which no target object is detected;
    在所述未检测到目标对象的视频帧的数量超过预设视频帧数量的情况下，检测到所述学习行为包括：在所述视频中的至少部分视频帧中未出现。In a case where the number of video frames in which the target object is not detected exceeds the preset number of video frames, detecting the learning behavior includes: not appearing in at least part of the video frames in the video.
  10. 根据权利要求1至9中任意一项所述的方法，其特征在于，所述根据所述目标对象执行所述至少一类学习行为的持续时间，生成学习状态信息，包括：The method according to any one of claims 1 to 9, wherein the generating learning state information according to the duration of the target object performing the at least one type of learning behavior comprises:
    在检测到所述目标对象执行至少一类学习行为的时间不小于时间阈值的情况下，记录至少一类所述学习行为的持续时间；If it is detected that the time for the target object to perform at least one type of learning behavior is not less than a time threshold, record the duration of at least one type of learning behavior;
    将至少一类所述学习行为对应的所述持续时间，作为所述学习状态信息。The duration corresponding to at least one type of the learning behavior is used as the learning state information.
  11. 根据权利要求1至10中任意一项所述的方法，其特征在于，所述方法还包括：The method according to any one of claims 1 to 10, wherein the method further comprises:
    对所述视频中的至少部分视频帧中的背景区域进行渲染，其中，所述背景区域为所述视频帧中所述目标对象以外的区域。Rendering a background area in at least a part of the video frame in the video, where the background area is an area outside the target object in the video frame.
  12. 根据权利要求1至11中任意一项所述的方法，其特征在于，所述方法还包括：The method according to any one of claims 1 to 11, wherein the method further comprises:
    统计至少一个所述目标对象的学习状态信息，得到至少一个所述目标对象的统计结果；Statistics the learning state information of at least one of the target objects, and obtain a statistical result of at least one of the target objects;
    根据至少一个所述目标对象的统计结果，生成学习状态统计数据。According to the statistical result of at least one of the target objects, the learning state statistical data is generated.
  13. 根据权利要求12所述的方法，其特征在于，所述根据至少一个所述目标对象的统计结果，生成学习状态统计数据，包括：The method according to claim 12, wherein said generating statistical data of learning state according to a statistical result of at least one of said target objects comprises:
    根据至少一个所述目标对象所属的类别，获取至少一个所述类别包含的目标对象的统计结果，生成至少一个类别的学习状态统计数据，其中，所述目标对象所属的类别包括所述目标对象参与的课程、所述目标对象注册的机构以及所述目标对象使用的设备中的至少一种；和/或，According to the category to which at least one of the target objects belongs, the statistical results of the target objects contained in at least one of the categories are obtained, and the learning status statistics data of at least one category are generated, wherein the category to which the target object belongs includes the participation of the target object At least one of the courses of, the institution registered by the target object, and the equipment used by the target object; and/or,
    将至少一个所述目标对象的统计结果进行可视化处理，生成至少一个所述目标对象的学习状态统计数据。Visual processing is performed on the statistical results of at least one of the target objects to generate statistical data of the learning state of at least one of the target objects.
  14. 一种视频处理装置，其特征在于，包括：A video processing device, characterized in that it comprises:
    视频获取模块，用于获取视频，其中，所述视频中的至少部分视频帧包含目标对象；A video acquisition module, configured to acquire a video, wherein at least part of the video frames in the video contain the target object;
    检测模块，用于根据所述视频，对所述目标对象在观看教学课程过程中的至少一类学习行为进行检测；The detection module is configured to detect at least one type of learning behavior of the target object in the process of watching the teaching course according to the video;
    生成模块，用于在检测到所述目标对象执行至少一类学习行为的情况下，根据至少部分包含所述至少一类学习行为的视频帧和/或所述目标对象执行所述至少一类学习行为的持续时间，生成学习状态信息。A generating module, configured to perform the at least one type of learning according to at least part of the video frame containing the at least one type of learning behavior and/or the target object in the case of detecting that the target object performs at least one type of learning behavior The duration of the behavior to generate learning status information.
  15. 一种电子设备，其特征在于，包括：An electronic device, characterized in that it comprises:
    处理器；processor;
    用于存储处理器可执行指令的存储器；A memory for storing processor executable instructions;
    其中，所述处理器被配置为调用所述存储器存储的指令，以执行权利要求1至12中任意一项所述的方法。Wherein, the processor is configured to call instructions stored in the memory to execute the method according to any one of claims 1-12.
  16. 一种计算机可读存储介质，其上存储有计算机程序指令，其特征在于，所述计算机程序指令被处理器执行时实现权利要求1至13中任意一项所述的方法。A computer-readable storage medium having computer program instructions stored thereon, wherein the computer program instructions implement the method according to any one of claims 1 to 13 when the computer program instructions are executed by a processor.
  17. 一种计算机程序，包括计算机可读代码，当所述计算机可读代码在电子设备中运行时，所述电子设备中的处理器执行用于实现权利要求1-13中的任一权利要求所述的方法。A computer program, comprising computer-readable code, when the computer-readable code runs in an electronic device, the processor in the electronic device executes the Methods.
PCT/CN2020/137690 2020-05-22 2020-12-18 Video processing method and apparatus, and electronic device and storage medium WO2021232775A1 (en)

Priority Applications (2)

Application Number Priority Date Filing Date Title
KR1020217021262A KR20210144658A (en) 2020-05-22 2020-12-18 Video processing method and apparatus, electronic device and storage medium
JP2021538705A JP2022537475A (en) 2020-05-22 2020-12-18 Video processing method and apparatus, electronic device and storage medium

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202010442733.6A CN111553323A (en) 2020-05-22 2020-05-22 Video processing method and device, electronic equipment and storage medium
CN202010442733.6 2020-05-22

Publications (1)

Publication Number Publication Date
WO2021232775A1 true WO2021232775A1 (en) 2021-11-25

Family

ID=72000950

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/137690 WO2021232775A1 (en) 2020-05-22 2020-12-18 Video processing method and apparatus, and electronic device and storage medium

Country Status (5)

Country Link
JP (1) JP2022537475A (en)
KR (1) KR20210144658A (en)
CN (1) CN111553323A (en)
TW (1) TW202145131A (en)
WO (1) WO2021232775A1 (en)

Families Citing this family (13)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111553323A (en) * 2020-05-22 2020-08-18 北京市商汤科技开发有限公司 Video processing method and device, electronic equipment and storage medium
CN112287844B (en) * 2020-10-30 2023-04-18 北京市商汤科技开发有限公司 Student situation analysis method and device, electronic device and storage medium
CN112652200A (en) * 2020-11-16 2021-04-13 北京家有课堂科技有限公司 Man-machine interaction system, man-machine interaction method, server, interaction control device and storage medium
TWI759016B (en) * 2020-12-17 2022-03-21 正文科技股份有限公司 Testee learning status detection method and testee learning status detection system
CN112598551B (en) * 2020-12-24 2022-11-29 北京市商汤科技开发有限公司 Behavior guidance scheme generation method and device, computer equipment and storage medium
CN112613780B (en) * 2020-12-29 2022-11-25 北京市商汤科技开发有限公司 Method and device for generating learning report, electronic equipment and storage medium
CN112866808B (en) * 2020-12-31 2022-09-06 北京市商汤科技开发有限公司 Video processing method and device, electronic equipment and storage medium
CN112990723B (en) * 2021-03-24 2021-11-30 食安快线信息技术(深圳)有限公司 Online education platform student learning force analysis feedback method based on user learning behavior deep analysis
CN113052088A (en) * 2021-03-29 2021-06-29 北京大米科技有限公司 Image processing method and device, readable storage medium and electronic equipment
CN114663261B (en) * 2022-05-18 2022-08-23 火焰蓝(浙江)信息科技有限公司 Data processing method suitable for training and examination system
CN114677751B (en) * 2022-05-26 2022-09-09 深圳市中文路教育科技有限公司 Learning state monitoring method, monitoring device and storage medium
CN116128453B (en) * 2023-02-18 2024-05-03 广州市点易资讯科技有限公司 Online course inspection method, system, equipment and medium
CN117636219B (en) * 2023-12-04 2024-06-14 浙江大学 Collaborative state analysis method and system in family sibling interaction process

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20160148515A1 (en) * 2014-11-20 2016-05-26 MyChild, Inc. Web and mobile parent engagement and learning management system
CN109815795A (en) * 2018-12-14 2019-05-28 深圳壹账通智能科技有限公司 Classroom student's state analysis method and device based on face monitoring
CN110033400A (en) * 2019-03-26 2019-07-19 深圳先进技术研究院 A kind of classroom monitoring analysis system
CN110991381A (en) * 2019-12-12 2020-04-10 山东大学 Real-time classroom student state analysis and indication reminding system and method based on behavior and voice intelligent recognition
CN111553323A (en) * 2020-05-22 2020-08-18 北京市商汤科技开发有限公司 Video processing method and device, electronic equipment and storage medium
CN112287844A (en) * 2020-10-30 2021-01-29 北京市商汤科技开发有限公司 Student situation analysis method and device, electronic device and storage medium

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2013097311A (en) * 2011-11-04 2013-05-20 Zenrin Datacom Co Ltd Learning support device, learning support method and learning support program
JPWO2018097177A1 (en) * 2016-11-24 2019-10-17 株式会社ガイア・システム・ソリューション Engagement measurement system
CN108399376B (en) * 2018-02-07 2020-11-06 华中师范大学 Intelligent analysis method and system for classroom learning interest of students
JP6636670B1 (en) * 2019-07-19 2020-01-29 株式会社フォーサイト Learning system, learning lecture providing method, and program

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20160148515A1 (en) * 2014-11-20 2016-05-26 MyChild, Inc. Web and mobile parent engagement and learning management system
CN109815795A (en) * 2018-12-14 2019-05-28 深圳壹账通智能科技有限公司 Classroom student's state analysis method and device based on face monitoring
CN110033400A (en) * 2019-03-26 2019-07-19 深圳先进技术研究院 A kind of classroom monitoring analysis system
CN110991381A (en) * 2019-12-12 2020-04-10 山东大学 Real-time classroom student state analysis and indication reminding system and method based on behavior and voice intelligent recognition
CN111553323A (en) * 2020-05-22 2020-08-18 北京市商汤科技开发有限公司 Video processing method and device, electronic equipment and storage medium
CN112287844A (en) * 2020-10-30 2021-01-29 北京市商汤科技开发有限公司 Student situation analysis method and device, electronic device and storage medium

Also Published As

Publication number Publication date
CN111553323A (en) 2020-08-18
KR20210144658A (en) 2021-11-30
JP2022537475A (en) 2022-08-26
TW202145131A (en) 2021-12-01

Similar Documents

Publication Publication Date Title
WO2021232775A1 (en) Video processing method and apparatus, and electronic device and storage medium
CN112287844B (en) Student situation analysis method and device, electronic device and storage medium
US9734410B2 (en) Systems and methods for analyzing facial expressions within an online classroom to gauge participant attentiveness
US20220319181A1 (en) Artificial intelligence (ai)-based system and method for managing education of students in real-time
CN111833861B (en) Event evaluation report generation based on artificial intelligence
US9955116B2 (en) Utilizing eye tracking to determine attendee engagement
US20230222932A1 (en) Methods, systems, and media for context-aware estimation of student attention in online learning
WO2021218194A1 (en) Data processing method and apparatus, electronic device, and storage medium
CN111353363A (en) Teaching effect detection method and device and electronic equipment
Zhao et al. Semi-automated 8 collaborative online training module for improving communication skills
CN111556279A (en) Monitoring method and communication method of instant session
CN110598632A (en) Target object monitoring method and device, electronic equipment and storage medium
Baecher Video in teacher learning: Through their own eyes
JP2009267621A (en) Communication apparatus
TWI528336B (en) Speech skills of audio and video automatic assessment and training system
CN116527828A (en) Image processing method and device, electronic equipment and readable storage medium
CN116912723A (en) Classroom broadcasting guiding method and device, electronic equipment and storage medium
CN114830208A (en) Information processing apparatus, information processing method, and program
CN111144255B (en) Analysis method and device for non-language behaviors of teacher
Kushalnagar et al. Improving classroom visual accessibility with cooperative smartphone recordings
Marino Implicitly Conveying Emotion While Teleconferencing
JP2022171465A (en) Program, method, and information processing device
CN115428466A (en) Simulating audience reaction to performers on a camera
CN115052194A (en) Learning report generation method, device, electronic equipment and storage medium

Legal Events

Date Code Title Description
ENP Entry into the national phase

Ref document number: 2021538705

Country of ref document: JP

Kind code of ref document: A

121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20937075

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20937075

Country of ref document: EP

Kind code of ref document: A1

WWE Wipo information: entry into national phase

Ref document number: 522432705

Country of ref document: SA

WWE Wipo information: entry into national phase

Ref document number: 522432705

Country of ref document: SA