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

WO2020168606A1 - Advertisement video optimising method, apparatus and device and computer readable storage medium - Google Patents

Advertisement video optimising method, apparatus and device and computer readable storage medium Download PDF

Info

Publication number
WO2020168606A1
WO2020168606A1 PCT/CN2019/078528 CN2019078528W WO2020168606A1 WO 2020168606 A1 WO2020168606 A1 WO 2020168606A1 CN 2019078528 W CN2019078528 W CN 2019078528W WO 2020168606 A1 WO2020168606 A1 WO 2020168606A1
Authority
WO
WIPO (PCT)
Prior art keywords
key
video frame
video
candidate
advertisement
Prior art date
Application number
PCT/CN2019/078528
Other languages
French (fr)
Chinese (zh)
Inventor
裴勇
郑文琛
杨强
Original Assignee
深圳前海微众银行股份有限公司
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 深圳前海微众银行股份有限公司 filed Critical 深圳前海微众银行股份有限公司
Publication of WO2020168606A1 publication Critical patent/WO2020168606A1/en

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/40—Client devices specifically adapted for the reception of or interaction with content, e.g. set-top-box [STB]; Operations thereof
    • H04N21/43—Processing of content or additional data, e.g. demultiplexing additional data from a digital video stream; Elementary client operations, e.g. monitoring of home network or synchronising decoder's clock; Client middleware
    • H04N21/44—Processing of video elementary streams, e.g. splicing a video clip retrieved from local storage with an incoming video stream or rendering scenes according to encoded video stream scene graphs
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/80—Generation or processing of content or additional data by content creator independently of the distribution process; Content per se
    • H04N21/81—Monomedia components thereof
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/80—Generation or processing of content or additional data by content creator independently of the distribution process; Content per se
    • H04N21/83—Generation or processing of protective or descriptive data associated with content; Content structuring
    • H04N21/84—Generation or processing of descriptive data, e.g. content descriptors
    • H04N21/8405—Generation or processing of descriptive data, e.g. content descriptors represented by keywords
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/80—Generation or processing of content or additional data by content creator independently of the distribution process; Content per se
    • H04N21/83—Generation or processing of protective or descriptive data associated with content; Content structuring
    • H04N21/845—Structuring of content, e.g. decomposing content into time segments

Definitions

  • the present invention relates to the technical field of the Internet, and in particular to an advertising video optimization method, device, equipment and computer-readable storage medium.
  • the advertising designer needs to reconsider the advertising optimization direction, and only start to produce the advertising video after the optimization direction is determined.
  • the production cycle of the advertising video is long, which requires more labor and time costs.
  • the optimization direction of the advertisement is artificially conceived by the advertisement designer, there is uncertainty, so that the optimized advertisement video still cannot achieve better delivery effects, and the advertisement video cannot be optimized quickly and accurately. Therefore, how to quickly and accurately optimize advertising videos is a problem that needs to be solved urgently.
  • the main purpose of the present invention is to provide an advertising video optimization method, device, equipment and computer readable storage medium, aiming at optimizing the advertising video quickly and accurately.
  • the advertising video optimization method includes the following steps:
  • the step of evaluating each candidate key video frame among the plurality of candidate key video frames according to the key index evaluation algorithm to obtain the target key index of each candidate key video frame includes:
  • the target key index of each candidate key video frame is determined.
  • the step of determining the target key index of each candidate key video frame includes:
  • the entropy of the text information and the similarity, the target key index of each candidate key video frame is determined.
  • the method further includes:
  • the entropy of the text information, and the aesthetic score, the target key index of each candidate key video frame is determined.
  • the method further includes:
  • the entropy of the text information, the similarity and the aesthetic score, the target key index of each candidate key video frame is determined.
  • the method further includes:
  • the algorithm strategy update instruction When the algorithm strategy update instruction is monitored, according to the algorithm strategy update instruction, the corresponding advertisement feedback data, the key index evaluation algorithm to be updated and the key video frame advance strategy are obtained;
  • the key index evaluation algorithm to be updated and the key video frame advance strategy are updated.
  • the steps of performing an update operation on the key index evaluation algorithm and the key video frame preamble strategy to be updated include:
  • the key index evaluation algorithm is updated according to the first parameter value, and the key video frame advance strategy is updated according to the second parameter value.
  • an advertisement video optimization device which includes:
  • the segment segmentation module is used to segment the advertisement video to be optimized into several advertisement video segments;
  • the video frame extraction module is used to extract key video frames for each advertisement video segment through a preset video frame extraction algorithm to obtain several candidate key video frames;
  • the key index evaluation module is used to obtain a key index evaluation algorithm, and according to the key index evaluation algorithm, evaluate each candidate key video frame among the plurality of candidate key video frames to obtain the target key index of each candidate key video frame ;
  • the selection module is configured to select a target key video frame from the plurality of candidate key video frames according to the target key index of each candidate key video frame;
  • the advertising video optimization module is configured to obtain a key video frame advance strategy, and based on the key video frame advance strategy and the target key video frame, optimize the advertisement video to obtain a target advertisement video.
  • the present invention also provides an advertisement video optimization device.
  • the advertisement video optimization device includes a memory, a processor, and an advertisement video optimization device stored in the memory and running on the processor.
  • a program when the advertisement video optimization program is executed by the processor, the following steps are implemented:
  • the present invention also provides a computer-readable storage medium on which an advertisement video optimization program is stored, and when the advertisement video optimization program is executed by a processor, the following steps are implemented:
  • the present invention provides an advertising video optimization method, device, equipment and computer readable storage medium.
  • the present invention divides the advertising video into several advertising video segments, and extracts key video frames for each advertising video segment to obtain several candidate key videos Then, based on the key index evaluation algorithm, each candidate key video frame is evaluated, and the target key index of each candidate key video frame is obtained. Based on the target key index of each candidate key video frame, from several candidate key videos Select the target key video frame in the frame, and finally optimize the advertising video based on the key video frame forward strategy and the target key video frame to obtain the target advertising video.
  • the whole process does not require the advertising designer to reconsider the optimization direction, which can avoid advertising
  • the designer’s uncertainty does not need to spend too much time to produce the advertising video.
  • the advertising video can be optimized quickly and accurately, and the advertising video can also be improved. Delivery effect.
  • FIG. 1 is a schematic diagram of a device structure of a hardware operating environment involved in a solution of an embodiment of the present invention
  • FIG. 2 is a schematic flowchart of the first embodiment of the advertising video optimization method of the present invention.
  • FIG. 3 is a schematic flowchart of a fourth embodiment of the advertising video optimization method of the present invention.
  • FIG. 4 is a schematic diagram of functional modules of the first embodiment of the advertising video optimization device of the present invention.
  • Fig. 1 is a schematic diagram of the device structure of the hardware operating environment involved in the solution of the embodiment of the present invention.
  • the advertising video optimization device in the embodiment of the present invention may be a PC, or a mobile terminal device with a display function, such as a smart phone, a tablet computer, or a portable computer.
  • the advertisement video optimization device may include a processor 1001, such as a CPU, a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005.
  • the communication bus 1002 is used to implement connection and communication between these components.
  • the user interface 1003 may include a display screen (Display) and an input unit such as a keyboard (Keyboard), and the optional user interface 1003 may also include a standard wired interface and a wireless interface.
  • the network interface 1004 may optionally include a standard wired interface and a wireless interface (such as a WI-FI interface).
  • the memory 1005 can be a high-speed RAM memory or a stable memory (non-volatile memory), such as disk storage.
  • the memory 1005 may also be a storage device independent of the foregoing processor 1001.
  • the structure of the advertisement video optimization device shown in FIG. 1 does not constitute a limitation on the advertisement video optimization device, and may include more or less components than those shown in the figure, or combine certain components, or be different.
  • the layout of the components does not constitute a limitation on the advertisement video optimization device, and may include more or less components than those shown in the figure, or combine certain components, or be different. The layout of the components.
  • a memory 1005 as a computer storage medium may include an operating system, a network communication module, a user interface module, and an advertisement video optimization program.
  • the network interface 1004 is mainly used to connect to the back-end server and communicate with the back-end server;
  • the user interface 1003 is mainly used to connect to the client (user side) and communicate with the client;
  • the processor 1001 may be used to call the advertisement video optimization program stored in the memory 1005 and execute the following steps:
  • processor 1001 may be used to call the advertisement video optimization program stored in the memory 1005, and further execute the following steps:
  • the target key index of each candidate key video frame is determined.
  • processor 1001 may be used to call the advertisement video optimization program stored in the memory 1005, and further execute the following steps:
  • the entropy of the text information and the similarity, the target key index of each candidate key video frame is determined.
  • processor 1001 may be used to call the advertisement video optimization program stored in the memory 1005, and further execute the following steps:
  • the entropy of the text information, and the aesthetic score, the target key index of each candidate key video frame is determined.
  • processor 1001 may be used to call the advertisement video optimization program stored in the memory 1005, and further execute the following steps:
  • the entropy of the text information, the similarity and the aesthetic score, the target key index of each candidate key video frame is determined.
  • processor 1001 may be used to call the advertisement video optimization program stored in the memory 1005, and further execute the following steps:
  • the algorithm strategy update instruction When the algorithm strategy update instruction is monitored, according to the algorithm strategy update instruction, the corresponding advertisement feedback data, the key index evaluation algorithm to be updated and the key video frame advance strategy are obtained;
  • the key index evaluation algorithm to be updated and the key video frame advance strategy are updated.
  • processor 1001 may be used to call the advertisement video optimization program stored in the memory 1005, and further execute the following steps:
  • the key index evaluation algorithm is updated according to the first parameter value, and the key video frame advance strategy is updated according to the second parameter value.
  • the specific embodiments of the advertising video optimization device of the present invention are basically the same as the specific embodiments of the following advertising video optimization methods, and will not be repeated here.
  • the invention provides an advertising video optimization method.
  • Fig. 2 is a schematic flowchart of a first embodiment of a method for optimizing an advertisement video according to the present invention.
  • the advertising video optimization method includes:
  • Step S101 dividing the advertisement video to be optimized into several advertisement video fragments, and extracting key video frames for each advertisement video fragment through a preset video frame extraction algorithm to obtain several candidate key video frames;
  • the advertising video optimization method is applied to an advertising video optimization device, and the advertising video optimization device can be the device shown in FIG. 1.
  • the advertising video optimization device is installed with an advertising video optimization program.
  • the advertisement video that needs to be optimized can be transmitted to the advertisement video optimization device.
  • the transmission methods include network transmission (such as transmitting the advertisement video to the advertisement video optimization device through a data network or wireless network) and local transmission (such as a data connection).
  • the advertisement video is transmitted to the advertisement video optimization device via cable, U disk or hard disk.
  • the user can select the advertisement video that needs to be optimized.
  • the advertisement video optimization device obtains the advertisement video to be optimized.
  • the advertisement video to be optimized is divided into several advertisement video fragments, and the key video frame of each advertisement video fragment is extracted through a preset video frame extraction algorithm to obtain several candidate key video frames.
  • the segmentation method of advertising video is Shot boundary detection (Shot boundary detection , SBD)
  • the advertisement video can be divided into several independent advertisement video fragments through Shot boundary detection, and each advertisement video fragment is composed of multiple video frames.
  • Shot boundary detection algorithms include, but are not limited to, color histograms based on video frames, brightness values based on video frames, and edge features based on video frames.
  • the preset video frame extraction algorithms include, but are not limited to, clustering algorithms and recurrent neural network algorithms. Specifically, the corresponding feature vector is extracted from each video frame, and based on the clustering algorithm or recurrent neural network algorithm and the extracted features Vector, extract key video frames for each advertisement video segment to obtain several candidate key video frames.
  • the clustering algorithm does not require a large amount of labeled video data for training, and is suitable for scenarios with a small number of samples or no labeled data
  • the recurrent neural network algorithm requires a large amount of labeled video data for training, which is suitable for a large number of samples and
  • different algorithms can be selected based on actual application scenarios.
  • Step S102 Obtain a key index evaluation algorithm, and evaluate each candidate key video frame among the plurality of candidate key video frames according to the key index evaluation algorithm to obtain a target key index of each candidate key video frame;
  • the advertising video optimization device after obtaining a number of candidate key video frames, obtains a key index evaluation algorithm, and according to the key index evaluation algorithm, evaluates each candidate key video frame among the plurality of candidate key video frames to obtain each The target key index of a candidate key video frame. Specifically, the text information contained in each candidate key video frame among several candidate key video frames is obtained, and the target key index of each candidate key video frame is determined according to the key index evaluation algorithm and the text information, that is, the target key index of each candidate key video frame is calculated.
  • the entropy (information amount) of the text information contained in the key video frame and query the mapping relationship table between the entropy of the text information and the key index stored in advance, and determine the key index corresponding to the entropy of the text information contained in each candidate key video frame as For the target key index of each candidate key video frame, it should be noted that the greater the entropy of the text information, the higher the key index, and the smaller the entropy of the text information, the lower the key index.
  • the specific method of obtaining the text information contained in the candidate key video frame is through OCR (Optical Character Recognition (optical character recognition) algorithm to obtain the text information contained in each candidate key video frame, that is, to perform denoising, binarization, tilt correction, character cutting and recognition on each candidate key video frame to obtain the candidate key video The text information in the frame.
  • OCR Optical Character Recognition
  • the specific method for determining the target key index of the candidate key video frame may also be to calculate the entropy of the text information contained in each candidate key video frame, and to calculate the difference between the text information contained in each candidate key video frame and the preset text information.
  • the entropy and similarity of the text information, the target key index of each candidate key video frame is determined, that is, the mapping relationship between the entropy of the pre-stored text information and the key index is searched to obtain each The first key index of the candidate key video frame, and query the pre-stored similarity and key index mapping table to obtain the second key index of each candidate key video frame, Then obtain the first weight coefficient and the second weight coefficient from the key index evaluation algorithm, and multiply the first weight coefficient of each candidate key video frame by the first weight coefficient to obtain the first weight key index of each candidate key video frame , Multiply the second key index of each candidate key video frame by the second weight coefficient to obtain the second weight key index of each candidate key video frame, and combine the first weight key index of each candidate key video frame with each The corresponding sum of the second weighted key indices of the candidate key video frames is determined as the target key index of each candidate key video frame. It should be noted that the sum of the first weight coefficient and the second weight coefficient
  • Step S103 according to the target key index of each candidate key video frame, select a target key video frame from the plurality of candidate key video frames;
  • the advertising video optimization device selects the target key video frame from a number of candidate key video frames according to the target key index of each candidate key video frame, that is, from Among several candidate key video frames, the candidate key video frame with the highest target key index is selected as the target key video frame.
  • Step S104 Obtain a key video frame advance strategy, and based on the key video frame advance strategy and the target key video frame, optimize the advertisement video to obtain a target advertisement video.
  • the advertising video optimization device after obtaining the target key video frame, obtains the key video frame pre-position strategy, and based on the key video frame pre-position strategy and the target key video frame, optimizes the advertisement video to obtain the target
  • the advertisement video that is, based on the key video frame advance strategy, advance the target key video frame at the corresponding position of the advertisement video, and set a corresponding play time for the target key video frame to optimize the advertisement video.
  • the key video frame advance strategy includes the length of the key video frame and the playing time of the video frame. It should be noted that the key video frame advance strategy can be set based on actual conditions, which is not specifically limited in this embodiment.
  • the present invention divides the advertising video into several advertising video segments, and extracts key video frames for each advertising video segment to obtain several candidate key video frames. Then, based on the key index evaluation algorithm, each candidate key video Frame evaluation to obtain the target key index of each candidate key video frame, and based on the target key index of each candidate key video frame, select the target key video frame from several candidate key video frames, and finally based on the key video frame front
  • the strategy and the target key video frame are optimized for the advertisement video to obtain the target advertisement video. The whole process does not require the advertisement designer to reconsider the optimization direction, which can avoid the advertisement designer’s uncertainty and does not require much time.
  • To make advertising videos, and optimize advertising videos based on key video frame extraction and key video frame pre-optimization can quickly and accurately optimize advertising videos, and also improve the delivery effect of advertising videos.
  • a second embodiment of the advertising video optimization method of the present invention is proposed.
  • the specific method for determining the target key index of each candidate key video frame can also be
  • the advertising video optimization device calculates the entropy of the text information contained in each candidate key video frame, and predicts each of several candidate key video frames based on a preset aesthetic score prediction model.
  • the aesthetic score of a candidate key video frame and then according to the key index evaluation algorithm, the entropy of the text information and the aesthetic score, the target key index of each candidate key video frame is determined, that is, the entropy of the pre-stored text information and the key index are determined
  • the mapping relationship table is used to obtain the first key index of each candidate key video frame, and the pre-stored mapping relationship table between aesthetic scores and key indexes is queried to obtain the third key index of each candidate key video frame, Then obtain the first weight coefficient and the third weight coefficient from the key index evaluation algorithm, and multiply the first key index of each candidate key video frame by the first weight coefficient to obtain the first weight key index of each candidate key video frame , Multiply the third key index of each candidate key video frame by the third weight coefficient to obtain the third weight key index of each candidate key video frame, and finally combine the first weight key index of each candidate key video frame with each
  • the corresponding sum of the third weighted key index of a candidate key video frame is determined as the target key index of each candidate key
  • each candidate key video frame is characterized By extraction, the aesthetic feature of each candidate key video frame required by the aesthetic score prediction model is obtained, and the aesthetic feature of each candidate key video frame is input to the aesthetic score prediction model to obtain the aesthetic score of each candidate key video frame.
  • the present invention comprehensively evaluates the candidate key video frames based on the entropy of the text information contained in the candidate key video frames and the aesthetic score of the candidate key video frames to obtain the target key index of the candidate key video frame, which can greatly improve the target
  • the credibility of the key index facilitates the subsequent accurate identification of the target key video frame.
  • a third embodiment of the advertising video optimization method of the present invention is proposed.
  • the difference from the previous embodiment is that after the aesthetic score of each candidate key video frame is predicted, the advertising video is optimized
  • the device calculates the similarity between the text information contained in each candidate key video frame and the preset text information, and determines the target key of each candidate key video frame according to the key index evaluation algorithm, the entropy of the text information, the similarity and the aesthetic score Index, that is, query the mapping relation table between the entropy of the pre-stored text information and the key index to obtain the first key index of each candidate key video frame, and query the pre-stored similarity and key index mapping table to obtain each candidate key The second key index of the video frame, And query the pre-stored mapping table of aesthetic scores and key indexes to obtain the third key index of each candidate key video frame, and then obtain the first weight coefficient, the second weight coefficient and the third weight coefficient from the key index evaluation algorithm, And multiply the first key index of each candidate key video frame
  • the present invention comprehensively evaluates the candidate key video frame based on the entropy of the text information contained in the candidate key video frame, the aesthetic score of the candidate key video frame, and the similarity between the text information contained in the candidate key video frame and the preset text information , To obtain the target key index of the candidate key video frame, which can further improve the credibility of the target key index, and facilitate the subsequent accurate finding of the target key video frame.
  • step S101 a fourth embodiment of the advertising video optimization method of the present invention is proposed.
  • the difference from the foregoing embodiment is that before step S101, it further includes:
  • Step S105 when the algorithm strategy update instruction is monitored, the corresponding advertisement feedback data, the key index evaluation algorithm to be updated, and the key video frame advance strategy are obtained according to the algorithm strategy update instruction;
  • Step S106 According to the advertisement feedback data, the key index evaluation algorithm and the key video frame advance strategy to be updated are updated.
  • the advertisement video optimization equipment updates the instruction according to the algorithm strategy, obtains the corresponding advertising feedback data, the key index evaluation algorithm to be updated, and the key video frame prepositioning strategy, and then according to the advertising feedback data, the key index evaluation algorithm and key video to be updated
  • the frame preamble strategy performs the update operation.
  • the advertisement feedback data includes, but is not limited to, click-through rate, user registration rate, brand reach rate, product purchase rate, and ad viewing time.
  • a preset algorithm strategy update model is obtained, and the advertisement feedback data is input to the algorithm strategy update model, and then the first parameter value output by the algorithm strategy update model for updating the key index evaluation algorithm and the value used for Update the second parameter value of the key video frame advance strategy, finally update the key index evaluation algorithm according to the first parameter value, and update the key video frame advance strategy according to the second parameter value.
  • the algorithm strategy update model is obtained through training of large amount of advertisement feedback data, first parameter value and second parameter value.
  • the specific training method can be set based on actual conditions, which is not specifically limited in this embodiment.
  • the first parameter value includes the value of the first weight coefficient, the value of the second weight coefficient, and the value of the third weight coefficient.
  • the second parameter value includes the value of the pre-length and the value of the video playback time, that is, the evaluation algorithm for the key index
  • the update is to update the value of the first weight coefficient, the value of the second weight coefficient, and the value of the third weight coefficient in the key index evaluation algorithm.
  • the update of the key video frame forward strategy is to update the key video The value of the leading length and the value of the video playing time in the frame leading strategy are updated.
  • the present invention updates the key index evaluation algorithm and the key video frame pre-positioning strategy, and adopts a better key index evaluation algorithm and key video frame pre-positioning strategy to accurately optimize the advertisement video. .
  • the invention also provides an advertisement video optimization device.
  • FIG. 4 is a schematic diagram of the functional modules of the first embodiment of the advertising video optimization device of the present invention.
  • the advertisement video optimization device includes:
  • the segment segmentation module 101 is configured to segment the advertisement video to be optimized into several advertisement video segments;
  • the video frame extraction module 102 is configured to extract key video frames for each advertisement video segment by using a preset video frame extraction algorithm to obtain several candidate key video frames;
  • the key index evaluation module 103 is configured to obtain a key index evaluation algorithm, and according to the key index evaluation algorithm, evaluate each candidate key video frame among the plurality of candidate key video frames to obtain the target key of each candidate key video frame index;
  • the selection module 104 is configured to select a target key video frame from the plurality of candidate key video frames according to the target key index of each candidate key video frame;
  • the advertisement video optimization module 105 is configured to obtain a key video frame advance strategy, and based on the key video frame advance strategy and the target key video frame, optimize the advertisement video to obtain a target advertisement video.
  • the key index evaluation module 103 is also used for:
  • the target key index of each candidate key video frame is determined.
  • the key index evaluation module 103 is also used for:
  • the entropy of the text information and the similarity, the target key index of each candidate key video frame is determined.
  • the key index evaluation module 103 is also used for:
  • the entropy of the text information, and the aesthetic score, the target key index of each candidate key video frame is determined.
  • the key index evaluation module 103 is also used for:
  • the entropy of the text information, the similarity and the aesthetic score, the target key index of each candidate key video frame is determined.
  • the advertisement video optimization device further includes:
  • the obtaining module is used to obtain corresponding advertisement feedback data, the key index evaluation algorithm to be updated, and the key video frame advance strategy according to the algorithm strategy update instruction when the algorithm strategy update instruction is monitored;
  • the update module is used to perform an update operation on the key index evaluation algorithm and key video frame advance strategy to be updated based on the advertisement feedback data.
  • update module is also used for:
  • the key index evaluation algorithm is updated according to the first parameter value, and the key video frame advance strategy is updated according to the second parameter value.
  • the specific embodiments of the advertisement video optimization device of the present invention are basically the same as the foregoing embodiments of the advertisement video optimization method, and will not be repeated here.
  • an embodiment of the present invention also provides a computer-readable storage medium having an advertisement video optimization program stored on the computer-readable storage medium, and when the advertisement video optimization program is executed by a processor, the following steps are performed:
  • the target key index of each candidate key video frame is determined.
  • the entropy of the text information and the similarity, the target key index of each candidate key video frame is determined.
  • the entropy of the text information, and the aesthetic score, the target key index of each candidate key video frame is determined.
  • the algorithm strategy update instruction When the algorithm strategy update instruction is monitored, according to the algorithm strategy update instruction, the corresponding advertisement feedback data, the key index evaluation algorithm to be updated and the key video frame advance strategy are obtained;
  • the key index evaluation algorithm to be updated and the key video frame advance strategy are updated.
  • the key index evaluation algorithm is updated according to the first parameter value, and the key video frame advance strategy is updated according to the second parameter value.
  • the specific embodiments of the computer-readable storage medium of the present invention are basically the same as the foregoing embodiments of the advertisement video optimization method, and will not be repeated here.
  • the technical solution of the present invention essentially or the part that contributes to the existing technology can be embodied in the form of a software product, and the computer software product is stored in a storage medium (such as ROM/RAM) as described above. , Magnetic disk, optical disk), including several instructions to make a terminal device (can be a mobile phone, computer, server, air conditioner, or network device, etc.) execute the method described in each embodiment of the present invention.

Landscapes

  • Engineering & Computer Science (AREA)
  • Multimedia (AREA)
  • Signal Processing (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)
  • Two-Way Televisions, Distribution Of Moving Picture Or The Like (AREA)

Abstract

Disclosed in the present invention are an advertisement video optimising method, apparatus, and device, and a computer readable storage medium, the method comprising: segmenting an advertisement video to be optimised into a plurality of advertisement video fragments and performing key video frame extraction on each advertisement video fragment by means of a preset video frame extraction algorithm to obtain a plurality of candidate key video frames; acquiring a key index evaluation algorithm and, on the basis of the key index evaluation algorithm, evaluating each candidate key video frame amongst the plurality of candidate key video frames to obtain a target key index of each candidate key video frame; on the basis of the target key index of each candidate key video frame, selecting a target key video frame from the plurality of candidate key video frames; acquiring a key video frame front-end policy and, on the basis of the key video frame front-end policy and the target key video frame, optimising the advertisement video to obtain a target advertisement video. The present invention can quickly and accurately optimise an advertisement video.

Description

广告视频优化方法、装置、设备及计算机可读存储介质 Advertising video optimization method, device, equipment and computer readable storage medium To
技术领域Technical field
本发明涉及互联网的技术领域，尤其涉及一种广告视频优化方法、装置、设备及计算机可读存储介质。The present invention relates to the technical field of the Internet, and in particular to an advertising video optimization method, device, equipment and computer-readable storage medium.
背景技术Background technique
随着金融科技（Fintech），尤其是互联网金融的快速发展，越来越多的企业逐渐采用线上的方式推广品牌和产品，可以通过各大广告平台投放用于推广品牌和产品的广告视频，可以让更多的人观看广告视频，便于大家了解品牌和产品，能够在用户需要产品时，用户可以通过相关的渠道购买产品。在投放广告视频的过程中，存在用户跳过了广告视频时，品牌和产品也还未能展示的问题，广告视频的投放效果较差，为此广告设计者需要优化广告视频。With the rapid development of financial technology (Fintech), especially Internet finance, more and more companies are gradually adopting online methods to promote brands and products, and they can use major advertising platforms to place advertising videos to promote brands and products. It can allow more people to watch the advertising video, so that everyone can understand the brand and product. When the user needs the product, the user can purchase the product through related channels. In the process of placing advertising videos, there is a problem that brands and products have not been displayed even when users skip the advertising videos. The effect of advertising videos is poor. For this reason, advertising designers need to optimize the advertising videos.
然而，在优化广告视频之前，广告设计者需要重新构思广告优化方向，在确定优化方向之后才开始制作广告视频，广告视频的制作周期较长，需要耗费较多的人力成本和时间成本，同时，由于广告的优化方向是广告设计者人为构思的，存在不确定性，使得优化后的广告视频仍不能取得较好的投放效果，无法快速准确的优化广告视频。因此，如何快速准确的优化广告视频是目前亟待解决的问题。However, before optimizing the advertising video, the advertising designer needs to reconsider the advertising optimization direction, and only start to produce the advertising video after the optimization direction is determined. The production cycle of the advertising video is long, which requires more labor and time costs. At the same time, Since the optimization direction of the advertisement is artificially conceived by the advertisement designer, there is uncertainty, so that the optimized advertisement video still cannot achieve better delivery effects, and the advertisement video cannot be optimized quickly and accurately. Therefore, how to quickly and accurately optimize advertising videos is a problem that needs to be solved urgently.
发明内容Summary of the invention
本发明的主要目的在于提供一种广告视频优化方法、装置、设备及计算机可读存储介质，旨在快速准确的优化广告视频。The main purpose of the present invention is to provide an advertising video optimization method, device, equipment and computer readable storage medium, aiming at optimizing the advertising video quickly and accurately.
为实现上述目的，本发明提供一种广告视频优化方法，所述广告视频优化方法包括以下步骤：In order to achieve the above objective, the present invention provides an advertising video optimization method. The advertising video optimization method includes the following steps:
将待优化的广告视频分割成若干广告视频片段，并通过预设的视频帧提取算法对每个广告视频片段进行关键视频帧提取，得到若干候选关键视频帧；Divide the advertisement video to be optimized into several advertisement video fragments, and extract key video frames from each advertisement video fragment through a preset video frame extraction algorithm to obtain several candidate key video frames;
获取关键指数评估算法，并依据所述关键指数评估算法，评估所述若干候选关键视频帧中的每一候选关键视频帧，得到每一候选关键视频帧的目标关键指数；Acquiring a key index evaluation algorithm, and evaluating each candidate key video frame among the plurality of candidate key video frames according to the key index evaluation algorithm, to obtain a target key index of each candidate key video frame;
依据所述每一候选关键视频帧的目标关键指数，从所述若干候选关键视频帧中选择目标关键视频帧；Selecting a target key video frame from the plurality of candidate key video frames according to the target key index of each candidate key video frame;
获取关键视频帧前置策略，并基于所述关键视频帧前置策略和所述目标关键视频帧，对所述广告视频进行优化，得到目标广告视频。Acquire a key video frame advance strategy, and based on the key video frame advance strategy and the target key video frame, optimize the advertisement video to obtain a target advertisement video.
进一步地，依据所述关键指数评估算法，评估所述若干候选关键视频帧中的每一候选关键视频帧，得到每一候选关键视频帧的目标关键指数的步骤包括：Further, the step of evaluating each candidate key video frame among the plurality of candidate key video frames according to the key index evaluation algorithm to obtain the target key index of each candidate key video frame includes:
获取所述若干候选关键视频帧中每一候选关键视频帧包含的文字信息；Acquiring text information contained in each candidate key video frame among the plurality of candidate key video frames;
依据所述关键指数评估算法和所述文字信息，确定每一候选关键视频帧的目标关键指数。According to the key index evaluation algorithm and the text information, the target key index of each candidate key video frame is determined.
进一步地，依据所述关键指数评估算法和所述文字信息，确定每一候选关键视频帧的目标关键指数的步骤包括：Further, according to the key index evaluation algorithm and the text information, the step of determining the target key index of each candidate key video frame includes:
计算所述每一候选关键视频帧包含的文字信息的熵，并计算所述每一候选关键视频帧包含的文字信息与预设文字信息的相似度；Calculating the entropy of the text information contained in each candidate key video frame, and calculating the similarity between the text information contained in each candidate key video frame and the preset text information;
依据所述关键指数评估算法、所述文字信息的熵和所述相似度，确定每一候选关键视频帧的目标关键指数。According to the key index evaluation algorithm, the entropy of the text information and the similarity, the target key index of each candidate key video frame is determined.
进一步地，获取所述若干候选关键视频帧中每一候选关键视频帧包含的文字信息的步骤之后，还包括：Further, after the step of obtaining the text information contained in each candidate key video frame among the plurality of candidate key video frames, the method further includes:
计算所述每一候选关键视频帧包含的文字信息的熵；Calculating the entropy of the text information contained in each candidate key video frame;
基于预设美学分数预测模型，预测所述若干候选关键视频帧中每一候选关键视频帧的美学分数；Predict the aesthetic score of each candidate key video frame among the plurality of candidate key video frames based on a preset aesthetic score prediction model;
依据所述关键指数评估算法、所述文字信息的熵和所述美学分数，确定所述每一候选关键视频帧的目标关键指数。According to the key index evaluation algorithm, the entropy of the text information, and the aesthetic score, the target key index of each candidate key video frame is determined.
进一步地，所述基于预设美学分数预测模型，预测所述若干候选关键视频帧中每一候选关键视频帧的美学分数的步骤之后，还包括：Further, after the step of predicting the aesthetic score of each candidate key video frame in the plurality of candidate key video frames based on a preset aesthetic score prediction model, the method further includes:
计算所述每一候选关键视频帧包含的文字信息与预设文字信息的相似度；Calculating the similarity between the text information contained in each candidate key video frame and the preset text information;
依据所述关键指数评估算法、所述文字信息的熵、所述相似度和所述美学分数，确定所述每一候选关键视频帧的目标关键指数。According to the key index evaluation algorithm, the entropy of the text information, the similarity and the aesthetic score, the target key index of each candidate key video frame is determined.
进一步地，所述将待优化的广告视频分割成若干广告视频片段的步骤之前，还包括：Further, before the step of dividing the advertisement video to be optimized into several advertisement video fragments, the method further includes:
当监测到算法策略更新指令时，依据所述算法策略更新指令，获取对应的广告反馈数据、待更新的关键指数评估算法和关键视频帧前置策略；When the algorithm strategy update instruction is monitored, according to the algorithm strategy update instruction, the corresponding advertisement feedback data, the key index evaluation algorithm to be updated and the key video frame advance strategy are obtained;
依据所述广告反馈数据，对待更新的关键指数评估算法和关键视频帧前置策略执行更新操作。According to the advertisement feedback data, the key index evaluation algorithm to be updated and the key video frame advance strategy are updated.
进一步地，依据所述广告反馈数据，对待更新的关键指数评估算法和关键视频帧前置策略执行更新操作的步骤包括：Further, according to the advertisement feedback data, the steps of performing an update operation on the key index evaluation algorithm and the key video frame preamble strategy to be updated include:
获取预设的算法策略更新模型，并将所述广告反馈数据输入至所述算法策略更新模型；Obtaining a preset algorithm strategy update model, and input the advertisement feedback data into the algorithm strategy update model;
获取所述算法策略更新模型输出的用于更新所述关键指数评估算法的第一参数值以及用于更新所述关键视频帧前置策略的第二参数值；Acquiring a first parameter value output by the algorithm strategy update model for updating the key index evaluation algorithm and a second parameter value for updating the key video frame forward strategy;
依据所述第一参数值更新所述关键指数评估算法，以及依据所述第二参数值更新所述关键视频帧前置策略。The key index evaluation algorithm is updated according to the first parameter value, and the key video frame advance strategy is updated according to the second parameter value.
此外，为实现上述目的，本发明还提供一种广告视频优化装置，所述广告视频优化装置包括：In addition, in order to achieve the above object, the present invention also provides an advertisement video optimization device, which includes:
片段分割模块，用于将待优化的广告视频分割成若干广告视频片段；The segment segmentation module is used to segment the advertisement video to be optimized into several advertisement video segments;
视频帧提取模块，用于通过预设的视频帧提取算法对每个广告视频片段进行关键视频帧提取，得到若干候选关键视频帧；The video frame extraction module is used to extract key video frames for each advertisement video segment through a preset video frame extraction algorithm to obtain several candidate key video frames;
关键指数评估模块，用于获取关键指数评估算法，并依据所述关键指数评估算法，评估所述若干候选关键视频帧中的每一候选关键视频帧，得到每一候选关键视频帧的目标关键指数；The key index evaluation module is used to obtain a key index evaluation algorithm, and according to the key index evaluation algorithm, evaluate each candidate key video frame among the plurality of candidate key video frames to obtain the target key index of each candidate key video frame ；
选择模块，用于依据所述每一候选关键视频帧的目标关键指数，从所述若干候选关键视频帧中选择目标关键视频帧；The selection module is configured to select a target key video frame from the plurality of candidate key video frames according to the target key index of each candidate key video frame;
广告视频优化模块，用于获取关键视频帧前置策略，并基于所述关键视频帧前置策略和所述目标关键视频帧，对所述广告视频进行优化，得到目标广告视频。The advertising video optimization module is configured to obtain a key video frame advance strategy, and based on the key video frame advance strategy and the target key video frame, optimize the advertisement video to obtain a target advertisement video.
此外，为实现上述目的，本发明还提供一种广告视频优化设备，所述广告视频优化设备包括：存储器、处理器及存储在所述存储器上并可在所述处理器上运行的广告视频优化程序，所述广告视频优化程序被所述处理器执行时实现以下步骤：In addition, in order to achieve the above-mentioned object, the present invention also provides an advertisement video optimization device. The advertisement video optimization device includes a memory, a processor, and an advertisement video optimization device stored in the memory and running on the processor. A program, when the advertisement video optimization program is executed by the processor, the following steps are implemented:
将待优化的广告视频分割成若干广告视频片段，并通过预设的视频帧提取算法对每个广告视频片段进行关键视频帧提取，得到若干候选关键视频帧；Divide the advertisement video to be optimized into several advertisement video fragments, and extract key video frames from each advertisement video fragment through a preset video frame extraction algorithm to obtain several candidate key video frames;
获取关键指数评估算法，并依据所述关键指数评估算法，评估所述若干候选关键视频帧中的每一候选关键视频帧，得到每一候选关键视频帧的目标关键指数；Acquiring a key index evaluation algorithm, and evaluating each candidate key video frame among the plurality of candidate key video frames according to the key index evaluation algorithm, to obtain a target key index of each candidate key video frame;
依据所述每一候选关键视频帧的目标关键指数，从所述若干候选关键视频帧中选择目标关键视频帧；Selecting a target key video frame from the plurality of candidate key video frames according to the target key index of each candidate key video frame;
获取关键视频帧前置策略，并基于所述关键视频帧前置策略和所述目标关键视频帧，对所述广告视频进行优化，得到目标广告视频。Acquire a key video frame advance strategy, and based on the key video frame advance strategy and the target key video frame, optimize the advertisement video to obtain a target advertisement video.
本发明还提供一种计算机可读存储介质，所述计算机可读存储介质上存储有广告视频优化程序，所述广告视频优化程序被处理器执行时实现以下步骤：The present invention also provides a computer-readable storage medium on which an advertisement video optimization program is stored, and when the advertisement video optimization program is executed by a processor, the following steps are implemented:
将待优化的广告视频分割成若干广告视频片段，并通过预设的视频帧提取算法对每个广告视频片段进行关键视频帧提取，得到若干候选关键视频帧；Divide the advertisement video to be optimized into several advertisement video fragments, and extract key video frames from each advertisement video fragment through a preset video frame extraction algorithm to obtain several candidate key video frames;
获取关键指数评估算法，并依据所述关键指数评估算法，评估所述若干候选关键视频帧中的每一候选关键视频帧，得到每一候选关键视频帧的目标关键指数；Acquiring a key index evaluation algorithm, and evaluating each candidate key video frame among the plurality of candidate key video frames according to the key index evaluation algorithm, to obtain a target key index of each candidate key video frame;
依据所述每一候选关键视频帧的目标关键指数，从所述若干候选关键视频帧中选择目标关键视频帧；Selecting a target key video frame from the plurality of candidate key video frames according to the target key index of each candidate key video frame;
获取关键视频帧前置策略，并基于所述关键视频帧前置策略和所述目标关键视频帧，对所述广告视频进行优化，得到目标广告视频。Acquire a key video frame advance strategy, and based on the key video frame advance strategy and the target key video frame, optimize the advertisement video to obtain a target advertisement video.
本发明提供一种广告视频优化方法、装置、设备及计算机可读存储介质，本发明将广告视频分割成若干广告视频片段，并对每个广告视频片段进行关键视频帧提取，得到若干候选关键视频帧，然后基于关键指数评估算法，对每一候选关键视频帧进行评估，得到每一候选关键视频帧的目标关键指数，并基于，每一候选关键视频帧的目标关键指数，从若干候选关键视频帧中选择目标关键视频帧，最后基于关键视频帧前置策略和该目标关键视频帧，对该广告视频进行优化，得到目标广告视频，整个过程不需要广告设计者重新构思优化方向，可以避免广告设计者的不确定性，也不需要耗费较多的时间去制作广告视频，同时基于关键视频帧提取和关键视频帧前置优化广告视频，能够快速准确的优化广告视频，也可以提高广告视频的投放效果。The present invention provides an advertising video optimization method, device, equipment and computer readable storage medium. The present invention divides the advertising video into several advertising video segments, and extracts key video frames for each advertising video segment to obtain several candidate key videos Then, based on the key index evaluation algorithm, each candidate key video frame is evaluated, and the target key index of each candidate key video frame is obtained. Based on the target key index of each candidate key video frame, from several candidate key videos Select the target key video frame in the frame, and finally optimize the advertising video based on the key video frame forward strategy and the target key video frame to obtain the target advertising video. The whole process does not require the advertising designer to reconsider the optimization direction, which can avoid advertising The designer’s uncertainty does not need to spend too much time to produce the advertising video. At the same time, based on the key video frame extraction and key video frame pre-optimization, the advertising video can be optimized quickly and accurately, and the advertising video can also be improved. Delivery effect.
附图说明Description of the drawings
图1是本发明实施例方案涉及的硬件运行环境的设备结构示意图；FIG. 1 is a schematic diagram of a device structure of a hardware operating environment involved in a solution of an embodiment of the present invention;
图2为本发明广告视频优化方法第一实施例的流程示意图；2 is a schematic flowchart of the first embodiment of the advertising video optimization method of the present invention;
图3为本发明广告视频优化方法第四实施例的流程示意图；3 is a schematic flowchart of a fourth embodiment of the advertising video optimization method of the present invention;
图4为本发明广告视频优化装置第一实施例的功能模块示意图。4 is a schematic diagram of functional modules of the first embodiment of the advertising video optimization device of the present invention.
本发明目的的实现、功能特点及优点将结合实施例，参照附图做进一步说明。The realization of the objectives, functional characteristics and advantages of the present invention will be further described in conjunction with the embodiments and with reference to the accompanying drawings.
具体实施方式detailed description
应当理解，此处所描述的具体实施例仅仅用以解释本发明，并不用于限定本发明。It should be understood that the specific embodiments described here are only used to explain the present invention, but not to limit the present invention.
如图1所示，图1是本发明实施例方案涉及的硬件运行环境的设备结构示意图。As shown in Fig. 1, Fig. 1 is a schematic diagram of the device structure of the hardware operating environment involved in the solution of the embodiment of the present invention.
本发明实施例广告视频优化设备可以是PC，也可以是智能手机、平板电脑、便携计算机等具有显示功能的可移动式终端设备。The advertising video optimization device in the embodiment of the present invention may be a PC, or a mobile terminal device with a display function, such as a smart phone, a tablet computer, or a portable computer.
如图1所示，该广告视频优化设备可以包括：处理器1001，例如CPU，通信总线1002，用户接口1003，网络接口1004，存储器1005。其中，通信总线1002用于实现这些组件之间的连接通信。用户接口1003可以包括显示屏（Display）、输入单元比如键盘（Keyboard），可选的用户接口1003还可以包括标准的有线接口、无线接口。网络接口1004可选的可以包括标准的有线接口、无线接口（如WI-FI接口）。存储器1005可以是高速RAM存储器，也可以是稳定的存储器（non-volatile memory），例如磁盘存储器。存储器1005可选的还可以是独立于前述处理器1001的存储装置。As shown in FIG. 1, the advertisement video optimization device may include a processor 1001, such as a CPU, a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. Among them, the communication bus 1002 is used to implement connection and communication between these components. The user interface 1003 may include a display screen (Display) and an input unit such as a keyboard (Keyboard), and the optional user interface 1003 may also include a standard wired interface and a wireless interface. The network interface 1004 may optionally include a standard wired interface and a wireless interface (such as a WI-FI interface). The memory 1005 can be a high-speed RAM memory or a stable memory (non-volatile memory), such as disk storage. Optionally, the memory 1005 may also be a storage device independent of the foregoing processor 1001.
本领域技术人员可以理解，图1中示出的广告视频优化设备结构并不构成对广告视频优化设备的限定，可以包括比图示更多或更少的部件，或者组合某些部件，或者不同的部件布置。Those skilled in the art can understand that the structure of the advertisement video optimization device shown in FIG. 1 does not constitute a limitation on the advertisement video optimization device, and may include more or less components than those shown in the figure, or combine certain components, or be different. The layout of the components.
如图1所示，作为一种计算机存储介质的存储器1005中可以包括操作系统、网络通信模块、用户接口模块以及广告视频优化程序。As shown in FIG. 1, a memory 1005 as a computer storage medium may include an operating system, a network communication module, a user interface module, and an advertisement video optimization program.
在图1所示的广告视频优化设备中，网络接口1004主要用于连接后台服务器，与后台服务器进行数据通信；用户接口1003主要用于连接客户端（用户端），与客户端进行数据通信；而处理器1001可以用于调用存储器1005中存储的广告视频优化程序，并执行以下步骤：In the advertising video optimization device shown in FIG. 1, the network interface 1004 is mainly used to connect to the back-end server and communicate with the back-end server; the user interface 1003 is mainly used to connect to the client (user side) and communicate with the client; The processor 1001 may be used to call the advertisement video optimization program stored in the memory 1005 and execute the following steps:
将待优化的广告视频分割成若干广告视频片段，并通过预设的视频帧提取算法对每个广告视频片段进行关键视频帧提取，得到若干候选关键视频帧；Divide the advertisement video to be optimized into several advertisement video fragments, and extract key video frames from each advertisement video fragment through a preset video frame extraction algorithm to obtain several candidate key video frames;
获取关键指数评估算法，并依据所述关键指数评估算法，评估所述若干候选关键视频帧中的每一候选关键视频帧，得到每一候选关键视频帧的目标关键指数；Acquiring a key index evaluation algorithm, and evaluating each candidate key video frame among the plurality of candidate key video frames according to the key index evaluation algorithm, to obtain a target key index of each candidate key video frame;
依据所述每一候选关键视频帧的目标关键指数，从所述若干候选关键视频帧中选择目标关键视频帧；Selecting a target key video frame from the plurality of candidate key video frames according to the target key index of each candidate key video frame;
获取关键视频帧前置策略，并基于所述关键视频帧前置策略和所述目标关键视频帧，对所述广告视频进行优化，得到目标广告视频。Acquire a key video frame advance strategy, and based on the key video frame advance strategy and the target key video frame, optimize the advertisement video to obtain a target advertisement video.
进一步地，处理器1001可以用于调用存储器1005中存储的广告视频优化程序，还执行以下步骤：Further, the processor 1001 may be used to call the advertisement video optimization program stored in the memory 1005, and further execute the following steps:
获取所述若干候选关键视频帧中每一候选关键视频帧包含的文字信息；Acquiring text information contained in each candidate key video frame among the plurality of candidate key video frames;
依据所述关键指数评估算法和所述文字信息，确定每一候选关键视频帧的目标关键指数。According to the key index evaluation algorithm and the text information, the target key index of each candidate key video frame is determined.
进一步地，处理器1001可以用于调用存储器1005中存储的广告视频优化程序，还执行以下步骤：Further, the processor 1001 may be used to call the advertisement video optimization program stored in the memory 1005, and further execute the following steps:
计算所述每一候选关键视频帧包含的文字信息的熵，并计算所述每一候选关键视频帧包含的文字信息与预设文字信息的相似度；Calculating the entropy of the text information contained in each candidate key video frame, and calculating the similarity between the text information contained in each candidate key video frame and the preset text information;
依据所述关键指数评估算法、所述文字信息的熵和所述相似度，确定每一候选关键视频帧的目标关键指数。According to the key index evaluation algorithm, the entropy of the text information and the similarity, the target key index of each candidate key video frame is determined.
进一步地，处理器1001可以用于调用存储器1005中存储的广告视频优化程序，还执行以下步骤：Further, the processor 1001 may be used to call the advertisement video optimization program stored in the memory 1005, and further execute the following steps:
计算所述每一候选关键视频帧包含的文字信息的熵；Calculating the entropy of the text information contained in each candidate key video frame;
基于预设美学分数预测模型，预测所述若干候选关键视频帧中每一候选关键视频帧的美学分数；Predict the aesthetic score of each candidate key video frame among the plurality of candidate key video frames based on a preset aesthetic score prediction model;
依据所述关键指数评估算法、所述文字信息的熵和所述美学分数，确定所述每一候选关键视频帧的目标关键指数。According to the key index evaluation algorithm, the entropy of the text information, and the aesthetic score, the target key index of each candidate key video frame is determined.
进一步地，处理器1001可以用于调用存储器1005中存储的广告视频优化程序，还执行以下步骤：Further, the processor 1001 may be used to call the advertisement video optimization program stored in the memory 1005, and further execute the following steps:
计算所述每一候选关键视频帧包含的文字信息与预设文字信息的相似度；Calculating the similarity between the text information contained in each candidate key video frame and the preset text information;
依据所述关键指数评估算法、所述文字信息的熵、所述相似度和所述美学分数，确定所述每一候选关键视频帧的目标关键指数。According to the key index evaluation algorithm, the entropy of the text information, the similarity and the aesthetic score, the target key index of each candidate key video frame is determined.
进一步地，处理器1001可以用于调用存储器1005中存储的广告视频优化程序，还执行以下步骤：Further, the processor 1001 may be used to call the advertisement video optimization program stored in the memory 1005, and further execute the following steps:
当监测到算法策略更新指令时，依据所述算法策略更新指令，获取对应的广告反馈数据、待更新的关键指数评估算法和关键视频帧前置策略；When the algorithm strategy update instruction is monitored, according to the algorithm strategy update instruction, the corresponding advertisement feedback data, the key index evaluation algorithm to be updated and the key video frame advance strategy are obtained;
依据所述广告反馈数据，对待更新的关键指数评估算法和关键视频帧前置策略执行更新操作。According to the advertisement feedback data, the key index evaluation algorithm to be updated and the key video frame advance strategy are updated.
进一步地，处理器1001可以用于调用存储器1005中存储的广告视频优化程序，还执行以下步骤：Further, the processor 1001 may be used to call the advertisement video optimization program stored in the memory 1005, and further execute the following steps:
获取预设的算法策略更新模型，并将所述广告反馈数据输入至所述算法策略更新模型；Obtaining a preset algorithm strategy update model, and input the advertisement feedback data into the algorithm strategy update model;
获取所述算法策略更新模型输出的用于更新所述关键指数评估算法的第一参数值以及用于更新所述关键视频帧前置策略的第二参数值；Acquiring a first parameter value output by the algorithm strategy update model for updating the key index evaluation algorithm and a second parameter value for updating the key video frame forward strategy;
依据所述第一参数值更新所述关键指数评估算法，以及依据所述第二参数值更新所述关键视频帧前置策略。The key index evaluation algorithm is updated according to the first parameter value, and the key video frame advance strategy is updated according to the second parameter value.
其中，本发明广告视频优化设备的具体实施例与下述广告视频优化方法的各具体实施例基本相同，在此不作赘述。Wherein, the specific embodiments of the advertising video optimization device of the present invention are basically the same as the specific embodiments of the following advertising video optimization methods, and will not be repeated here.
本发明提供一种广告视频优化方法。The invention provides an advertising video optimization method.
参照图2，图2为本发明广告视频优化方法第一实施例的流程示意图。Referring to Fig. 2, Fig. 2 is a schematic flowchart of a first embodiment of a method for optimizing an advertisement video according to the present invention.
本实施例中，该广告视频优化方法包括：In this embodiment, the advertising video optimization method includes:
步骤S101，将待优化的广告视频分割成若干广告视频片段，并通过预设的视频帧提取算法对每个广告视频片段进行关键视频帧提取，得到若干候选关键视频帧；Step S101, dividing the advertisement video to be optimized into several advertisement video fragments, and extracting key video frames for each advertisement video fragment through a preset video frame extraction algorithm to obtain several candidate key video frames;
本实施例中，该广告视频优化方法应用于广告视频优化设备，该广告视频优化设备可选为图1所示的设备，该广告视频优化设备中安装有广告视频优化程序，当需要优化广告视频时，可以将需要优化的广告视频传输到广告视频优化设备中，传输的方式包括联网传输（如通过数据网络或无线网络将广告视频传输到广告视频优化设备中）和本地传输（如通过数据连接线、U盘或硬盘等将广告视频传输到广告视频优化设备中），用户可以选择需要优化的广告视频，当用户选择需要优化的广告视频之后，该广告视频优化设备获取待优化的广告视频，将待优化的广告视频分割成若干广告视频片段，并通过预设的视频帧提取算法对每个广告视频片段进行关键视频帧提取，得到若干候选关键视频帧。In this embodiment, the advertising video optimization method is applied to an advertising video optimization device, and the advertising video optimization device can be the device shown in FIG. 1. The advertising video optimization device is installed with an advertising video optimization program. At the time, the advertisement video that needs to be optimized can be transmitted to the advertisement video optimization device. The transmission methods include network transmission (such as transmitting the advertisement video to the advertisement video optimization device through a data network or wireless network) and local transmission (such as a data connection The advertisement video is transmitted to the advertisement video optimization device via cable, U disk or hard disk. The user can select the advertisement video that needs to be optimized. After the user selects the advertisement video that needs to be optimized, the advertisement video optimization device obtains the advertisement video to be optimized. The advertisement video to be optimized is divided into several advertisement video fragments, and the key video frame of each advertisement video fragment is extracted through a preset video frame extraction algorithm to obtain several candidate key video frames.
其中，广告视频的分割方式为Shot边界检测（Shot boundary detection ，SBD），通过Shot边界检测可以将广告视频分割成若干独立的广告视频片段，且每个广告视频片段由多个视频帧组成。Shot边界检测算法包括但不限于基于视频帧的颜色直方图、基于视频帧的亮度值和基于视频帧的边缘特征。预设的视频帧提取算法包括但不限于聚类算法和递归神经网络算法，具体地，从每一视频帧中提取对应的特征向量，并基于聚类算法或递归神经网络算法和提取出来的特征向量，对每个广告视频片段进行关键视频帧提取，得到若干候选关键视频帧。其中，聚类算法不需要大量有标注的视频数据进行训练，适合于样本数较少或没有标注数据的场景，而递归神经网络算法需要大量有标注的视频数据进行训练，适合样本数较多且有标注数据的场景，可基于实际的应用场景选择不同的算法。Among them, the segmentation method of advertising video is Shot boundary detection (Shot boundary detection , SBD), the advertisement video can be divided into several independent advertisement video fragments through Shot boundary detection, and each advertisement video fragment is composed of multiple video frames. Shot boundary detection algorithms include, but are not limited to, color histograms based on video frames, brightness values based on video frames, and edge features based on video frames. The preset video frame extraction algorithms include, but are not limited to, clustering algorithms and recurrent neural network algorithms. Specifically, the corresponding feature vector is extracted from each video frame, and based on the clustering algorithm or recurrent neural network algorithm and the extracted features Vector, extract key video frames for each advertisement video segment to obtain several candidate key video frames. Among them, the clustering algorithm does not require a large amount of labeled video data for training, and is suitable for scenarios with a small number of samples or no labeled data, while the recurrent neural network algorithm requires a large amount of labeled video data for training, which is suitable for a large number of samples and For scenarios with labeled data, different algorithms can be selected based on actual application scenarios.
步骤S102，获取关键指数评估算法，并依据所述关键指数评估算法，评估所述若干候选关键视频帧中的每一候选关键视频帧，得到每一候选关键视频帧的目标关键指数；Step S102: Obtain a key index evaluation algorithm, and evaluate each candidate key video frame among the plurality of candidate key video frames according to the key index evaluation algorithm to obtain a target key index of each candidate key video frame;
本实施例中，在得到若干候选关键视频帧之后，该广告视频优化设备获取关键指数评估算法，并依据该关键指数评估算法，评估若干候选关键视频帧中的每一候选关键视频帧，得到每一候选关键视频帧的目标关键指数。具体地，获取若干候选关键视频帧中每一候选关键视频帧包含的文字信息，并依据该关键指数评估算法和该文字信息，确定每一候选关键视频帧的目标关键指数，即计算每一候选关键视频帧包含的文字信息的熵（信息量），并查询预存的文字信息的熵与关键指数的映射关系表，将每一候选关键视频帧包含的文字信息的熵对应的关键指数，确定为每一候选关键视频帧的目标关键指数，需要说明的是，文字信息的熵越大，则关键指数越高，文字信息的熵越小，则关键指数越低。其中，候选关键视频帧包含的文字信息的具体获取方式为通过OCR （Optical Character Recognition，光学字符识别）算法，获取每一候选关键视频帧包含的文字信息，即对每一候选关键视频帧进行去噪、二值化、倾斜矫正、字符切割和识别等处理，获取候选关键视频帧中的文字信息。In this embodiment, after obtaining a number of candidate key video frames, the advertising video optimization device obtains a key index evaluation algorithm, and according to the key index evaluation algorithm, evaluates each candidate key video frame among the plurality of candidate key video frames to obtain each The target key index of a candidate key video frame. Specifically, the text information contained in each candidate key video frame among several candidate key video frames is obtained, and the target key index of each candidate key video frame is determined according to the key index evaluation algorithm and the text information, that is, the target key index of each candidate key video frame is calculated. The entropy (information amount) of the text information contained in the key video frame, and query the mapping relationship table between the entropy of the text information and the key index stored in advance, and determine the key index corresponding to the entropy of the text information contained in each candidate key video frame as For the target key index of each candidate key video frame, it should be noted that the greater the entropy of the text information, the higher the key index, and the smaller the entropy of the text information, the lower the key index. Among them, the specific method of obtaining the text information contained in the candidate key video frame is through OCR (Optical Character Recognition (optical character recognition) algorithm to obtain the text information contained in each candidate key video frame, that is, to perform denoising, binarization, tilt correction, character cutting and recognition on each candidate key video frame to obtain the candidate key video The text information in the frame.
具体地，候选关键视频帧的目标关键指数的具体确定方式还可以为计算每一候选关键视频帧包含的文字信息的熵，并计算每一候选关键视频帧包含的文字信息与预设文字信息的相似度，然后依据该关键指数评估算法、文字信息的熵和相似度，确定每一候选关键视频帧的目标关键指数，即查询预存的文字信息的熵与关键指数的映射关系表，得到每一候选关键视频帧的第一关键指数，以及查询预存的相似度与关键指数的映射关系表，得到每一候选关键视频帧的第二关键指数， 然后从关键指数评估算法中获取第一权重系数和第二权重系数，并用每一候选关键视频帧的第一关键指数乘以第一权重系数，得到每一候选关键视频帧的第一权重关键指数，用每一候选关键视频帧的第二关键指数乘以第二权重系数，得到每一候选关键视频帧的第二权重关键指数，将每一候选关键视频帧的第一权重关键指数与每一候选关键视频帧的第二权重关键指数的对应之和，确定为每一候选关键视频帧的目标关键指数。需要说明的是，第一权重系数和第二权重系数之和为1，且预设文字信息可基于实际情况进行设置，本实施例对此不作具体限定。Specifically, the specific method for determining the target key index of the candidate key video frame may also be to calculate the entropy of the text information contained in each candidate key video frame, and to calculate the difference between the text information contained in each candidate key video frame and the preset text information. Similarity, and then according to the key index evaluation algorithm, the entropy and similarity of the text information, the target key index of each candidate key video frame is determined, that is, the mapping relationship between the entropy of the pre-stored text information and the key index is searched to obtain each The first key index of the candidate key video frame, and query the pre-stored similarity and key index mapping table to obtain the second key index of each candidate key video frame, Then obtain the first weight coefficient and the second weight coefficient from the key index evaluation algorithm, and multiply the first weight coefficient of each candidate key video frame by the first weight coefficient to obtain the first weight key index of each candidate key video frame , Multiply the second key index of each candidate key video frame by the second weight coefficient to obtain the second weight key index of each candidate key video frame, and combine the first weight key index of each candidate key video frame with each The corresponding sum of the second weighted key indices of the candidate key video frames is determined as the target key index of each candidate key video frame. It should be noted that the sum of the first weight coefficient and the second weight coefficient is 1, and the preset text information can be set based on actual conditions, which is not specifically limited in this embodiment.
步骤S103，依据所述每一候选关键视频帧的目标关键指数，从所述若干候选关键视频帧中选择目标关键视频帧；Step S103, according to the target key index of each candidate key video frame, select a target key video frame from the plurality of candidate key video frames;
本实施例中，在得到每一候选关键视频帧的关键指数之后，该广告视频优化设备依据每一候选关键视频帧的目标关键指数，从若干候选关键视频帧中选择目标关键视频帧，即从若干候选关键视频帧中选择目标关键指数最高的候选关键视频帧作为目标关键视频帧。In this embodiment, after obtaining the key index of each candidate key video frame, the advertising video optimization device selects the target key video frame from a number of candidate key video frames according to the target key index of each candidate key video frame, that is, from Among several candidate key video frames, the candidate key video frame with the highest target key index is selected as the target key video frame.
步骤S104，获取关键视频帧前置策略，并基于所述关键视频帧前置策略和所述目标关键视频帧，对所述广告视频进行优化，得到目标广告视频。Step S104: Obtain a key video frame advance strategy, and based on the key video frame advance strategy and the target key video frame, optimize the advertisement video to obtain a target advertisement video.
本实施例中，在得到目标关键视频帧之后，该广告视频优化设备获取关键视频帧前置策略，并基于该关键视频帧前置策略和该目标关键视频帧，对广告视频进行优化，得到目标广告视频，即基于该关键视频帧前置策略，在该广告视频的对应位置前置该目标关键视频帧，且给该目标关键视频帧设置对应的播放时间，以优化广告视频。其中，该关键视频帧前置策略包含前置长度和视频帧的播放时间，需要说明的是，该关键视频帧前置策略可基于实际情况进行设置，本实施例对此不作具体限定。In this embodiment, after obtaining the target key video frame, the advertising video optimization device obtains the key video frame pre-position strategy, and based on the key video frame pre-position strategy and the target key video frame, optimizes the advertisement video to obtain the target The advertisement video, that is, based on the key video frame advance strategy, advance the target key video frame at the corresponding position of the advertisement video, and set a corresponding play time for the target key video frame to optimize the advertisement video. The key video frame advance strategy includes the length of the key video frame and the playing time of the video frame. It should be noted that the key video frame advance strategy can be set based on actual conditions, which is not specifically limited in this embodiment.
本实施例中，本发明将广告视频分割成若干广告视频片段，并对每个广告视频片段进行关键视频帧提取，得到若干候选关键视频帧，然后基于关键指数评估算法，对每一候选关键视频帧进行评估，得到每一候选关键视频帧的目标关键指数，并基于，每一候选关键视频帧的目标关键指数，从若干候选关键视频帧中选择目标关键视频帧，最后基于关键视频帧前置策略和该目标关键视频帧，对该广告视频进行优化，得到目标广告视频，整个过程不需要广告设计者重新构思优化方向，可以避免广告设计者的不确定性，也不需要耗费较多的时间去制作广告视频，同时基于关键视频帧提取和关键视频帧前置优化广告视频，能够快速准确的优化广告视频，也可以提高广告视频的投放效果。In this embodiment, the present invention divides the advertising video into several advertising video segments, and extracts key video frames for each advertising video segment to obtain several candidate key video frames. Then, based on the key index evaluation algorithm, each candidate key video Frame evaluation to obtain the target key index of each candidate key video frame, and based on the target key index of each candidate key video frame, select the target key video frame from several candidate key video frames, and finally based on the key video frame front The strategy and the target key video frame are optimized for the advertisement video to obtain the target advertisement video. The whole process does not require the advertisement designer to reconsider the optimization direction, which can avoid the advertisement designer’s uncertainty and does not require much time. To make advertising videos, and optimize advertising videos based on key video frame extraction and key video frame pre-optimization can quickly and accurately optimize advertising videos, and also improve the delivery effect of advertising videos.
进一步地，基于上述第一实施例，提出了本发明广告视频优化方法的第二实施例，与前述实施例的区别在于，每一候选关键视频帧的目标关键指数的具体确定方式还可以为在获取到每一候选关键视频帧包含的文字信息之后，该广告视频优化设备计算每一候选关键视频帧包含的文字信息的熵，并基于预设美学分数预测模型，预测若干候选关键视频帧中每一候选关键视频帧的美学分数，然后依据该关键指数评估算法、文字信息的熵和该美学分数，确定每一候选关键视频帧的目标关键指数，即查询预存的文字信息的熵与关键指数的映射关系表，得到每一候选关键视频帧的第一关键指数，以及查询预存的美学分数与关键指数的映射关系表，得到每一候选关键视频帧的第三关键指数， 然后从关键指数评估算法中获取第一权重系数和第三权重系数，并用每一候选关键视频帧的第一关键指数乘以第一权重系数，得到每一候选关键视频帧的第一权重关键指数，用每一候选关键视频帧的第三关键指数乘以第三权重系数，得到每一候选关键视频帧的第三权重关键指数，最后将每一候选关键视频帧的第一权重关键指数与每一候选关键视频帧的第三权重关键指数的对应之和，确定为每一候选关键视频帧的目标关键指数。需要说明的是，第一权重系数和第三权重系数之和为1，且第一权重系数和第三权重系数的具体值可基于实际情况设置，本实施例对此不作具体限定。Further, based on the above-mentioned first embodiment, a second embodiment of the advertising video optimization method of the present invention is proposed. The difference from the previous embodiment is that the specific method for determining the target key index of each candidate key video frame can also be After acquiring the text information contained in each candidate key video frame, the advertising video optimization device calculates the entropy of the text information contained in each candidate key video frame, and predicts each of several candidate key video frames based on a preset aesthetic score prediction model. The aesthetic score of a candidate key video frame, and then according to the key index evaluation algorithm, the entropy of the text information and the aesthetic score, the target key index of each candidate key video frame is determined, that is, the entropy of the pre-stored text information and the key index are determined The mapping relationship table is used to obtain the first key index of each candidate key video frame, and the pre-stored mapping relationship table between aesthetic scores and key indexes is queried to obtain the third key index of each candidate key video frame, Then obtain the first weight coefficient and the third weight coefficient from the key index evaluation algorithm, and multiply the first key index of each candidate key video frame by the first weight coefficient to obtain the first weight key index of each candidate key video frame , Multiply the third key index of each candidate key video frame by the third weight coefficient to obtain the third weight key index of each candidate key video frame, and finally combine the first weight key index of each candidate key video frame with each The corresponding sum of the third weighted key index of a candidate key video frame is determined as the target key index of each candidate key video frame. It should be noted that the sum of the first weighting coefficient and the third weighting coefficient is 1, and the specific values of the first weighting coefficient and the third weighting coefficient can be set based on actual conditions, which is not specifically limited in this embodiment.
其中，通过大数据量的视频样本数据可以训练得到美学分数预测模型，该美学分数预测模型可基于实际情况设置，本实施例对此不作具体限定，具体地，对每一候选关键视频帧进行特征提取，得到美学分数预测模型所需的每一候选关键视频帧的美学特征，并将每一候选关键视频帧的美学特征输入至美学分数预测模型，可以得到每一候选关键视频帧的美学分数。Among them, a large amount of video sample data can be trained to obtain an aesthetic score prediction model. The aesthetic score prediction model can be set based on actual conditions. This embodiment does not specifically limit this. Specifically, each candidate key video frame is characterized By extraction, the aesthetic feature of each candidate key video frame required by the aesthetic score prediction model is obtained, and the aesthetic feature of each candidate key video frame is input to the aesthetic score prediction model to obtain the aesthetic score of each candidate key video frame.
本实施例中，本发明基于候选关键视频帧包含的文字信息的熵和候选关键视频帧的美学分数，综合评估候选关键视频帧，得到候选关键视频帧的目标关键指数，能够极大的提高目标关键指数的可信度，便于后续准确地找出目标关键视频帧。In this embodiment, the present invention comprehensively evaluates the candidate key video frames based on the entropy of the text information contained in the candidate key video frames and the aesthetic score of the candidate key video frames to obtain the target key index of the candidate key video frame, which can greatly improve the target The credibility of the key index facilitates the subsequent accurate identification of the target key video frame.
进一步地，基于上述第二实施例，提出了本发明广告视频优化方法的第三实施例，与前述实施例的区别在于，在预测得到每一候选关键视频帧的美学分数之后，该广告视频优化设备计算每一候选关键视频帧包含的文字信息与预设文字信息的相似度，并依据该关键指数评估算法、文字信息的熵、相似度和美学分数，确定每一候选关键视频帧的目标关键指数，即查询预存的文字信息的熵与关键指数的映射关系表，得到每一候选关键视频帧的第一关键指数，以及查询预存的相似度与关键指数的映射关系表，得到每一候选关键视频帧的第二关键指数， 以及查询预存的美学分数与关键指数的映射关系表，得到每一候选关键视频帧的第三关键指数，然后从关键指数评估算法中获取第一权重系数、第二权重系数和第三权重系数，并用每一候选关键视频帧的第一关键指数乘以第一权重系数，得到每一候选关键视频帧的第一权重关键指数，用每一候选关键视频帧的第二关键指数乘以第二权重系数，得到每一候选关键视频帧的第二权重关键指数，以及用每一候选关键视频帧的第三关键指数乘以第三权重系数，得到每一候选关键视频帧的第三权重关键指数，最后将每一候选关键视频帧的第一权重关键指数、每一候选关键视频帧的第二权重关键指数与每一候选关键视频帧的第三权重关键指数之和，确定为每一候选关键视频帧的目标关键指数。需要说明的是，第一权重系数、第二权重系数和第三权重系数之和为1，且第一权重系数、第二权重系数和第三权重系数的具体值可基于实际情况进行设置，本实施例对此不作具体限定。Further, based on the above-mentioned second embodiment, a third embodiment of the advertising video optimization method of the present invention is proposed. The difference from the previous embodiment is that after the aesthetic score of each candidate key video frame is predicted, the advertising video is optimized The device calculates the similarity between the text information contained in each candidate key video frame and the preset text information, and determines the target key of each candidate key video frame according to the key index evaluation algorithm, the entropy of the text information, the similarity and the aesthetic score Index, that is, query the mapping relation table between the entropy of the pre-stored text information and the key index to obtain the first key index of each candidate key video frame, and query the pre-stored similarity and key index mapping table to obtain each candidate key The second key index of the video frame, And query the pre-stored mapping table of aesthetic scores and key indexes to obtain the third key index of each candidate key video frame, and then obtain the first weight coefficient, the second weight coefficient and the third weight coefficient from the key index evaluation algorithm, And multiply the first key index of each candidate key video frame by the first weight coefficient to obtain the first weight key index of each candidate key video frame, and multiply the second key index of each candidate key video frame by the second weight Coefficient, the second weighted key index of each candidate key video frame is obtained, and the third key index of each candidate key video frame is multiplied by the third weight coefficient to obtain the third weighted key index of each candidate key video frame, Finally, the sum of the first weight key index of each candidate key video frame, the second weight key index of each candidate key video frame and the third weight key index of each candidate key video frame is determined as each candidate key video The target key index of the frame. It should be noted that the sum of the first weight coefficient, the second weight coefficient, and the third weight coefficient is 1, and the specific values of the first weight coefficient, the second weight coefficient, and the third weight coefficient can be set based on actual conditions. The embodiment does not specifically limit this.
本实施例中，本发明基于候选关键视频帧包含的文字信息的熵、候选关键视频帧的美学分数和候选关键视频帧包含的文字信息与预设文字信息的相似度，综合评估候选关键视频帧，得到候选关键视频帧的目标关键指数，能够进一步地提高目标关键指数的可信度，便于后续准确地找出目标关键视频帧。In this embodiment, the present invention comprehensively evaluates the candidate key video frame based on the entropy of the text information contained in the candidate key video frame, the aesthetic score of the candidate key video frame, and the similarity between the text information contained in the candidate key video frame and the preset text information , To obtain the target key index of the candidate key video frame, which can further improve the credibility of the target key index, and facilitate the subsequent accurate finding of the target key video frame.
进一步地，参照图3，基于上述第一、第二或第三实施例，提出了本发明广告视频优化方法的第四实施例，与前述实施例的区别在于，该步骤S101之前，还包括：Further, referring to FIG. 3, based on the above-mentioned first, second or third embodiment, a fourth embodiment of the advertising video optimization method of the present invention is proposed. The difference from the foregoing embodiment is that before step S101, it further includes:
步骤S105，当监测到算法策略更新指令时，依据所述算法策略更新指令，获取对应的广告反馈数据、待更新的关键指数评估算法和关键视频帧前置策略；Step S105, when the algorithm strategy update instruction is monitored, the corresponding advertisement feedback data, the key index evaluation algorithm to be updated, and the key video frame advance strategy are obtained according to the algorithm strategy update instruction;
步骤S106，依据所述广告反馈数据，对待更新的关键指数评估算法和关键视频帧前置策略执行更新操作。Step S106: According to the advertisement feedback data, the key index evaluation algorithm and the key video frame advance strategy to be updated are updated.
本实施例中，在将待优化的广告视频分割成若干广告视频片段之前，存在需要更新关键指数评估算法和关键视频帧前置策略的情况，为此，当监测到算法策略更新指令时，该广告视频优化设备依据该算法策略更新指令，获取对应的广告反馈数据、待更新的关键指数评估算法和关键视频帧前置策略，然后依据该广告反馈数据，对待更新的关键指数评估算法和关键视频帧前置策略执行更新操作。其中，广告反馈数据包括但不限于点击率、用户注册率、品牌触达率、产品购买率和广告观看时长。In this embodiment, before the advertisement video to be optimized is divided into several advertisement video fragments, there are situations where it is necessary to update the key index evaluation algorithm and the key video frame preamble strategy. For this reason, when the algorithm strategy update instruction is detected, the The advertising video optimization equipment updates the instruction according to the algorithm strategy, obtains the corresponding advertising feedback data, the key index evaluation algorithm to be updated, and the key video frame prepositioning strategy, and then according to the advertising feedback data, the key index evaluation algorithm and key video to be updated The frame preamble strategy performs the update operation. Among them, the advertisement feedback data includes, but is not limited to, click-through rate, user registration rate, brand reach rate, product purchase rate, and ad viewing time.
具体地，获取预设的算法策略更新模型，并将该广告反馈数据输入至算法策略更新模型，然后获取该算法策略更新模型输出的用于更新该关键指数评估算法的第一参数值以及用于更新该关键视频帧前置策略的第二参数值，最后依据该第一参数值更新关键指数评估算法，以及依据第二参数值更新该关键视频帧前置策略。其中，该算法策略更新模型为通过大数据量的广告反馈数据、第一参数值和第二参数值训练得到的，具体的训练方式可基于实际情况进行设置，本实施例对此不作具体限定，且第一参数值包括第一权重系数的值、第二权重系数的值和第三权重系数的值，第二参数值包括前置长度的值和视频播放时间的值，即对关键指数评估算法进行更新，则是对关键指数评估算法中的第一权重系数的值、第二权重系数的值和第三权重系数的值进行更新，对关键视频帧前置策略进行更新，则是对关键视频帧前置策略中的前置长度的值和视频播放时间的值进行更新。Specifically, a preset algorithm strategy update model is obtained, and the advertisement feedback data is input to the algorithm strategy update model, and then the first parameter value output by the algorithm strategy update model for updating the key index evaluation algorithm and the value used for Update the second parameter value of the key video frame advance strategy, finally update the key index evaluation algorithm according to the first parameter value, and update the key video frame advance strategy according to the second parameter value. Wherein, the algorithm strategy update model is obtained through training of large amount of advertisement feedback data, first parameter value and second parameter value. The specific training method can be set based on actual conditions, which is not specifically limited in this embodiment. And the first parameter value includes the value of the first weight coefficient, the value of the second weight coefficient, and the value of the third weight coefficient. The second parameter value includes the value of the pre-length and the value of the video playback time, that is, the evaluation algorithm for the key index The update is to update the value of the first weight coefficient, the value of the second weight coefficient, and the value of the third weight coefficient in the key index evaluation algorithm. The update of the key video frame forward strategy is to update the key video The value of the leading length and the value of the video playing time in the frame leading strategy are updated.
本实施例中，本发明在优化广告视频之前，对关键指数评估算法和关键视频帧前置策略进行更新，采用更优的关键指数评估算法和关键视频帧前置策略，可以准确的优化广告视频。In this embodiment, before optimizing the advertisement video, the present invention updates the key index evaluation algorithm and the key video frame pre-positioning strategy, and adopts a better key index evaluation algorithm and key video frame pre-positioning strategy to accurately optimize the advertisement video. .
本发明还提供一种广告视频优化装置。The invention also provides an advertisement video optimization device.
参照图4，图4为本发明广告视频优化装置第一实施例的功能模块示意图。Referring to FIG. 4, FIG. 4 is a schematic diagram of the functional modules of the first embodiment of the advertising video optimization device of the present invention.
本实施例中，该广告视频优化装置包括：In this embodiment, the advertisement video optimization device includes:
片段分割模块101，用于将待优化的广告视频分割成若干广告视频片段；The segment segmentation module 101 is configured to segment the advertisement video to be optimized into several advertisement video segments;
视频帧提取模块102，用于通过预设的视频帧提取算法对每个广告视频片段进行关键视频帧提取，得到若干候选关键视频帧；The video frame extraction module 102 is configured to extract key video frames for each advertisement video segment by using a preset video frame extraction algorithm to obtain several candidate key video frames;
关键指数评估模块103，用于获取关键指数评估算法，并依据所述关键指数评估算法，评估所述若干候选关键视频帧中的每一候选关键视频帧，得到每一候选关键视频帧的目标关键指数；The key index evaluation module 103 is configured to obtain a key index evaluation algorithm, and according to the key index evaluation algorithm, evaluate each candidate key video frame among the plurality of candidate key video frames to obtain the target key of each candidate key video frame index;
选择模块104，用于依据所述每一候选关键视频帧的目标关键指数，从所述若干候选关键视频帧中选择目标关键视频帧；The selection module 104 is configured to select a target key video frame from the plurality of candidate key video frames according to the target key index of each candidate key video frame;
广告视频优化模块105，用于获取关键视频帧前置策略，并基于所述关键视频帧前置策略和所述目标关键视频帧，对所述广告视频进行优化，得到目标广告视频。The advertisement video optimization module 105 is configured to obtain a key video frame advance strategy, and based on the key video frame advance strategy and the target key video frame, optimize the advertisement video to obtain a target advertisement video.
进一步地，所述关键指数评估模块103还用于：Further, the key index evaluation module 103 is also used for:
获取所述若干候选关键视频帧中每一候选关键视频帧包含的文字信息；Acquiring text information contained in each candidate key video frame among the plurality of candidate key video frames;
依据所述关键指数评估算法和所述文字信息，确定每一候选关键视频帧的目标关键指数。According to the key index evaluation algorithm and the text information, the target key index of each candidate key video frame is determined.
进一步地，所述关键指数评估模块103还用于：Further, the key index evaluation module 103 is also used for:
计算所述每一候选关键视频帧包含的文字信息的熵，并计算所述每一候选关键视频帧包含的文字信息与预设文字信息的相似度；Calculating the entropy of the text information contained in each candidate key video frame, and calculating the similarity between the text information contained in each candidate key video frame and the preset text information;
依据所述关键指数评估算法、所述文字信息的熵和所述相似度，确定每一候选关键视频帧的目标关键指数。According to the key index evaluation algorithm, the entropy of the text information and the similarity, the target key index of each candidate key video frame is determined.
进一步地，所述关键指数评估模块103还用于：Further, the key index evaluation module 103 is also used for:
计算所述每一候选关键视频帧包含的文字信息的熵；Calculating the entropy of the text information contained in each candidate key video frame;
基于预设美学分数预测模型，预测所述若干候选关键视频帧中每一候选关键视频帧的美学分数；Predict the aesthetic score of each candidate key video frame among the plurality of candidate key video frames based on a preset aesthetic score prediction model;
依据所述关键指数评估算法、所述文字信息的熵和所述美学分数，确定所述每一候选关键视频帧的目标关键指数。According to the key index evaluation algorithm, the entropy of the text information, and the aesthetic score, the target key index of each candidate key video frame is determined.
进一步地，所述关键指数评估模块103还用于：Further, the key index evaluation module 103 is also used for:
计算所述每一候选关键视频帧包含的文字信息与预设文字信息的相似度；Calculating the similarity between the text information contained in each candidate key video frame and the preset text information;
依据所述关键指数评估算法、所述文字信息的熵、所述相似度和所述美学分数，确定所述每一候选关键视频帧的目标关键指数。According to the key index evaluation algorithm, the entropy of the text information, the similarity and the aesthetic score, the target key index of each candidate key video frame is determined.
进一步地，所述广告视频优化装置还包括：Further, the advertisement video optimization device further includes:
获取模块，用于当监测到算法策略更新指令时，依据所述算法策略更新指令，获取对应的广告反馈数据、待更新的关键指数评估算法和关键视频帧前置策略；The obtaining module is used to obtain corresponding advertisement feedback data, the key index evaluation algorithm to be updated, and the key video frame advance strategy according to the algorithm strategy update instruction when the algorithm strategy update instruction is monitored;
更新模块，用于依据所述广告反馈数据，对待更新的关键指数评估算法和关键视频帧前置策略执行更新操作。The update module is used to perform an update operation on the key index evaluation algorithm and key video frame advance strategy to be updated based on the advertisement feedback data.
进一步地，所述更新模块还用于：Further, the update module is also used for:
获取预设的算法策略更新模型，并将所述广告反馈数据输入至所述算法策略更新模型；Obtaining a preset algorithm strategy update model, and input the advertisement feedback data into the algorithm strategy update model;
获取所述算法策略更新模型输出的用于更新所述关键指数评估算法的第一参数值以及用于更新所述关键视频帧前置策略的第二参数值；Acquiring a first parameter value output by the algorithm strategy update model for updating the key index evaluation algorithm and a second parameter value for updating the key video frame forward strategy;
依据所述第一参数值更新所述关键指数评估算法，以及依据所述第二参数值更新所述关键视频帧前置策略。The key index evaluation algorithm is updated according to the first parameter value, and the key video frame advance strategy is updated according to the second parameter value.
其中，本发明广告视频优化装置的具体实施例与上述广告视频优化方法各实施例基本相同，在此不作赘述。Among them, the specific embodiments of the advertisement video optimization device of the present invention are basically the same as the foregoing embodiments of the advertisement video optimization method, and will not be repeated here.
此外，本发明实施例还提出一种计算机可读存储介质，所述计算机可读存储介质上存储有广告视频优化程序，所述广告视频优化程序被处理器执行时，执行以下步骤：In addition, an embodiment of the present invention also provides a computer-readable storage medium having an advertisement video optimization program stored on the computer-readable storage medium, and when the advertisement video optimization program is executed by a processor, the following steps are performed:
将待优化的广告视频分割成若干广告视频片段，并通过预设的视频帧提取算法对每个广告视频片段进行关键视频帧提取，得到若干候选关键视频帧；Divide the advertisement video to be optimized into several advertisement video fragments, and extract key video frames from each advertisement video fragment through a preset video frame extraction algorithm to obtain several candidate key video frames;
获取关键指数评估算法，并依据所述关键指数评估算法，评估所述若干候选关键视频帧中的每一候选关键视频帧，得到每一候选关键视频帧的目标关键指数；Acquiring a key index evaluation algorithm, and evaluating each candidate key video frame among the plurality of candidate key video frames according to the key index evaluation algorithm, to obtain a target key index of each candidate key video frame;
依据所述每一候选关键视频帧的目标关键指数，从所述若干候选关键视频帧中选择目标关键视频帧；Selecting a target key video frame from the plurality of candidate key video frames according to the target key index of each candidate key video frame;
获取关键视频帧前置策略，并基于所述关键视频帧前置策略和所述目标关键视频帧，对所述广告视频进行优化，得到目标广告视频。Acquire a key video frame advance strategy, and based on the key video frame advance strategy and the target key video frame, optimize the advertisement video to obtain a target advertisement video.
进一步地，所述广告视频优化程序被处理器执行时，还执行以下步骤：Further, when the advertisement video optimization program is executed by the processor, the following steps are also executed:
获取所述若干候选关键视频帧中每一候选关键视频帧包含的文字信息；Acquiring text information contained in each candidate key video frame among the plurality of candidate key video frames;
依据所述关键指数评估算法和所述文字信息，确定每一候选关键视频帧的目标关键指数。According to the key index evaluation algorithm and the text information, the target key index of each candidate key video frame is determined.
进一步地，所述广告视频优化程序被处理器执行时，还执行以下步骤：Further, when the advertisement video optimization program is executed by the processor, the following steps are also executed:
计算所述每一候选关键视频帧包含的文字信息的熵，并计算所述每一候选关键视频帧包含的文字信息与预设文字信息的相似度；Calculating the entropy of the text information contained in each candidate key video frame, and calculating the similarity between the text information contained in each candidate key video frame and the preset text information;
依据所述关键指数评估算法、所述文字信息的熵和所述相似度，确定每一候选关键视频帧的目标关键指数。According to the key index evaluation algorithm, the entropy of the text information and the similarity, the target key index of each candidate key video frame is determined.
进一步地，所述广告视频优化程序被处理器执行时，还执行以下步骤：Further, when the advertisement video optimization program is executed by the processor, the following steps are also executed:
计算所述每一候选关键视频帧包含的文字信息的熵；Calculating the entropy of the text information contained in each candidate key video frame;
基于预设美学分数预测模型，预测所述若干候选关键视频帧中每一候选关键视频帧的美学分数；Predict the aesthetic score of each candidate key video frame among the plurality of candidate key video frames based on a preset aesthetic score prediction model;
依据所述关键指数评估算法、所述文字信息的熵和所述美学分数，确定所述每一候选关键视频帧的目标关键指数。According to the key index evaluation algorithm, the entropy of the text information, and the aesthetic score, the target key index of each candidate key video frame is determined.
进一步地，所述广告视频优化程序被处理器执行时，还执行以下步骤：Further, when the advertisement video optimization program is executed by the processor, the following steps are also executed:
计算所述每一候选关键视频帧包含的文字信息与预设文字信息的相似度；Calculating the similarity between the text information contained in each candidate key video frame and the preset text information;
依据所述关键指数评估算法、所述文字信息的熵、所述相似度和所述美学分数，确定所述每一候选关键视频帧的目标关键指数Determine the target key index of each candidate key video frame according to the key index evaluation algorithm, the entropy of the text information, the similarity and the aesthetic score
进一步地，所述广告视频优化程序被处理器执行时，还执行以下步骤：Further, when the advertisement video optimization program is executed by the processor, the following steps are also executed:
当监测到算法策略更新指令时，依据所述算法策略更新指令，获取对应的广告反馈数据、待更新的关键指数评估算法和关键视频帧前置策略；When the algorithm strategy update instruction is monitored, according to the algorithm strategy update instruction, the corresponding advertisement feedback data, the key index evaluation algorithm to be updated and the key video frame advance strategy are obtained;
依据所述广告反馈数据，对待更新的关键指数评估算法和关键视频帧前置策略执行更新操作。According to the advertisement feedback data, the key index evaluation algorithm to be updated and the key video frame advance strategy are updated.
进一步地，所述广告视频优化程序被处理器执行时，还执行以下步骤：Further, when the advertisement video optimization program is executed by the processor, the following steps are also executed:
获取预设的算法策略更新模型，并将所述广告反馈数据输入至所述算法策略更新模型；Obtaining a preset algorithm strategy update model, and input the advertisement feedback data into the algorithm strategy update model;
获取所述算法策略更新模型输出的用于更新所述关键指数评估算法的第一参数值以及用于更新所述关键视频帧前置策略的第二参数值；Acquiring a first parameter value output by the algorithm strategy update model for updating the key index evaluation algorithm and a second parameter value for updating the key video frame forward strategy;
依据所述第一参数值更新所述关键指数评估算法，以及依据所述第二参数值更新所述关键视频帧前置策略。The key index evaluation algorithm is updated according to the first parameter value, and the key video frame advance strategy is updated according to the second parameter value.
其中，本发明计算机可读存储介质的具体实施例与上述广告视频优化方法各实施例基本相同，在此不作赘述。Among them, the specific embodiments of the computer-readable storage medium of the present invention are basically the same as the foregoing embodiments of the advertisement video optimization method, and will not be repeated here.
需要说明的是，在本文中，术语“包括”、“包含”或者其任何其他变体意在涵盖非排他性的包含，从而使得包括一系列要素的过程、方法、物品或者系统不仅包括那些要素，而且还包括没有明确列出的其他要素，或者是还包括为这种过程、方法、物品或者系统所固有的要素。在没有更多限制的情况下，由语句“包括一个……”限定的要素，并不排除在包括该要素的过程、方法、物品或者系统中还存在另外的相同要素。It should be noted that in this article, the terms "include", "include" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or system including a series of elements not only includes those elements, It also includes other elements that are not explicitly listed, or elements inherent to the process, method, article, or system. If there are no more restrictions, the element defined by the sentence "including a..." does not exclude the existence of other identical elements in the process, method, article or system that includes the element.
上述本发明实施例序号仅仅为了描述，不代表实施例的优劣。The sequence numbers of the foregoing embodiments of the present invention are only for description, and do not represent the superiority of the embodiments.
通过以上的实施方式的描述，本领域的技术人员可以清楚地了解到上述实施例方法可借助软件加必需的通用硬件平台的方式来实现，当然也可以通过硬件，但很多情况下前者是更佳的实施方式。基于这样的理解，本发明的技术方案本质上或者说对现有技术做出贡献的部分可以以软件产品的形式体现出来，该计算机软件产品存储在如上所述的一个存储介质(如ROM/RAM、磁碟、光盘)中，包括若干指令用以使得一台终端设备(可以是手机，计算机，服务器，空调器，或者网络设备等)执行本发明各个实施例所述的方法。Through the description of the above embodiments, those skilled in the art can clearly understand that the method of the above embodiments can be implemented by means of software plus the necessary general hardware platform. Of course, it can also be achieved by hardware, but in many cases the former is better.的实施方式。 Based on this understanding, the technical solution of the present invention essentially or the part that contributes to the existing technology can be embodied in the form of a software product, and the computer software product is stored in a storage medium (such as ROM/RAM) as described above. , Magnetic disk, optical disk), including several instructions to make a terminal device (can be a mobile phone, computer, server, air conditioner, or network device, etc.) execute the method described in each embodiment of the present invention.
以上仅为本发明的优选实施例，并非因此限制本发明的专利范围，凡是利用本发明说明书及附图内容所作的等效结构或等效流程变换，或直接或间接运用在其他相关的技术领域，均同理包括在本发明的专利保护范围内。The above are only the preferred embodiments of the present invention, and do not limit the scope of the present invention. Any equivalent structure or equivalent process transformation made by using the contents of the description and drawings of the present invention, or directly or indirectly applied to other related technical fields , The same reason is included in the scope of patent protection of the present invention.

Claims (20)

  1. 一种广告视频优化方法，其特征在于，所述广告视频优化方法包括以下步骤： An advertising video optimization method, characterized in that the advertising video optimization method includes the following steps:
    将待优化的广告视频分割成若干广告视频片段，并通过预设的视频帧提取算法对每个广告视频片段进行关键视频帧提取，得到若干候选关键视频帧；Divide the advertisement video to be optimized into several advertisement video fragments, and extract key video frames from each advertisement video fragment through a preset video frame extraction algorithm to obtain several candidate key video frames;
    获取关键指数评估算法，并依据所述关键指数评估算法，评估所述若干候选关键视频帧中的每一候选关键视频帧，得到每一候选关键视频帧的目标关键指数；Acquiring a key index evaluation algorithm, and evaluating each candidate key video frame among the plurality of candidate key video frames according to the key index evaluation algorithm, to obtain a target key index of each candidate key video frame;
    依据所述每一候选关键视频帧的目标关键指数，从所述若干候选关键视频帧中选择目标关键视频帧；Selecting a target key video frame from the plurality of candidate key video frames according to the target key index of each candidate key video frame;
    获取关键视频帧前置策略，并基于所述关键视频帧前置策略和所述目标关键视频帧，对所述广告视频进行优化，得到目标广告视频。Acquire a key video frame advance strategy, and based on the key video frame advance strategy and the target key video frame, optimize the advertisement video to obtain a target advertisement video.
  2. 如权利要求1所述的广告视频优化方法，其特征在于，依据所述关键指数评估算法，评估所述若干候选关键视频帧中的每一候选关键视频帧，得到每一候选关键视频帧的目标关键指数的步骤包括：The advertising video optimization method of claim 1, wherein each candidate key video frame of the plurality of candidate key video frames is evaluated according to the key index evaluation algorithm to obtain the target of each candidate key video frame The key index steps include:
    获取所述若干候选关键视频帧中每一候选关键视频帧包含的文字信息；Acquiring text information contained in each candidate key video frame among the plurality of candidate key video frames;
    依据所述关键指数评估算法和所述文字信息，确定每一候选关键视频帧的目标关键指数。According to the key index evaluation algorithm and the text information, the target key index of each candidate key video frame is determined.
  3. 如权利要求2所述的广告视频优化方法，其特征在于，依据所述关键指数评估算法和所述文字信息，确定每一候选关键视频帧的目标关键指数的步骤包括：3. The advertising video optimization method of claim 2, wherein the step of determining the target key index of each candidate key video frame according to the key index evaluation algorithm and the text information comprises:
    计算所述每一候选关键视频帧包含的文字信息的熵，并计算所述每一候选关键视频帧包含的文字信息与预设文字信息的相似度；Calculating the entropy of the text information contained in each candidate key video frame, and calculating the similarity between the text information contained in each candidate key video frame and the preset text information;
    依据所述关键指数评估算法、所述文字信息的熵和所述相似度，确定每一候选关键视频帧的目标关键指数。According to the key index evaluation algorithm, the entropy of the text information and the similarity, the target key index of each candidate key video frame is determined.
  4. 如权利要求2所述的广告视频优化方法，其特征在于，获取所述若干候选关键视频帧中每一候选关键视频帧包含的文字信息的步骤之后，还包括：3. The advertising video optimization method of claim 2, wherein after the step of obtaining the text information contained in each of the plurality of candidate key video frames, the method further comprises:
    计算所述每一候选关键视频帧包含的文字信息的熵；Calculating the entropy of the text information contained in each candidate key video frame;
    基于预设美学分数预测模型，预测所述若干候选关键视频帧中每一候选关键视频帧的美学分数；Predict the aesthetic score of each candidate key video frame among the plurality of candidate key video frames based on a preset aesthetic score prediction model;
    依据所述关键指数评估算法、所述文字信息的熵和所述美学分数，确定所述每一候选关键视频帧的目标关键指数。According to the key index evaluation algorithm, the entropy of the text information, and the aesthetic score, the target key index of each candidate key video frame is determined.
  5. 如权利要求4所述的广告视频优化方法，其特征在于，所述基于预设美学分数预测模型，预测所述若干候选关键视频帧中每一候选关键视频帧的美学分数的步骤之后，还包括：5. The advertising video optimization method of claim 4, wherein after the step of predicting the aesthetic score of each candidate key video frame in the plurality of candidate key video frames based on a preset aesthetic score prediction model, the method further comprises :
    计算所述每一候选关键视频帧包含的文字信息与预设文字信息的相似度；Calculating the similarity between the text information contained in each candidate key video frame and the preset text information;
    依据所述关键指数评估算法、所述文字信息的熵、所述相似度和所述美学分数，确定所述每一候选关键视频帧的目标关键指数。According to the key index evaluation algorithm, the entropy of the text information, the similarity and the aesthetic score, the target key index of each candidate key video frame is determined.
  6. 如权利要求1所述的广告视频优化方法，其特征在于，所述将待优化的广告视频分割成若干广告视频片段的步骤之前，还包括：5. The method for optimizing advertisement video according to claim 1, wherein before the step of dividing the advertisement video to be optimized into several advertisement video segments, the method further comprises:
    当监测到算法策略更新指令时，依据所述算法策略更新指令，获取对应的广告反馈数据、待更新的关键指数评估算法和关键视频帧前置策略；When the algorithm strategy update instruction is monitored, according to the algorithm strategy update instruction, the corresponding advertisement feedback data, the key index evaluation algorithm to be updated and the key video frame advance strategy are obtained;
    依据所述广告反馈数据，对待更新的关键指数评估算法和关键视频帧前置策略执行更新操作。According to the advertisement feedback data, the key index evaluation algorithm to be updated and the key video frame advance strategy are updated.
  7. 如权利要求6所述的广告视频优化方法，其特征在于，依据所述广告反馈数据，对待更新的关键指数评估算法和关键视频帧前置策略执行更新操作的步骤包括：8. The advertising video optimization method of claim 6, wherein the step of performing an update operation on the key index evaluation algorithm to be updated and the key video frame advance strategy based on the advertising feedback data comprises:
    获取预设的算法策略更新模型，并将所述广告反馈数据输入至所述算法策略更新模型；Obtaining a preset algorithm strategy update model, and input the advertisement feedback data into the algorithm strategy update model;
    获取所述算法策略更新模型输出的用于更新所述关键指数评估算法的第一参数值以及用于更新所述关键视频帧前置策略的第二参数值；Acquiring a first parameter value output by the algorithm strategy update model for updating the key index evaluation algorithm and a second parameter value for updating the key video frame forward strategy;
    依据所述第一参数值更新所述关键指数评估算法，以及依据所述第二参数值更新所述关键视频帧前置策略。The key index evaluation algorithm is updated according to the first parameter value, and the key video frame advance strategy is updated according to the second parameter value.
  8. 一种广告视频优化装置，其特征在于，所述广告视频优化装置包括：An advertisement video optimization device, characterized in that, the advertisement video optimization device includes:
    片段分割模块，用于将待优化的广告视频分割成若干广告视频片段；The segment segmentation module is used to segment the advertisement video to be optimized into several advertisement video segments;
    视频帧提取模块，用于通过预设的视频帧提取算法对每个广告视频片段进行关键视频帧提取，得到若干候选关键视频帧；The video frame extraction module is used to extract key video frames for each advertisement video segment through a preset video frame extraction algorithm to obtain several candidate key video frames;
    关键指数评估模块，用于获取关键指数评估算法，并依据所述关键指数评估算法，评估所述若干候选关键视频帧中的每一候选关键视频帧，得到每一候选关键视频帧的目标关键指数；The key index evaluation module is used to obtain a key index evaluation algorithm, and according to the key index evaluation algorithm, evaluate each candidate key video frame among the plurality of candidate key video frames to obtain the target key index of each candidate key video frame ；
    选择模块，用于依据所述每一候选关键视频帧的目标关键指数，从所述若干候选关键视频帧中选择目标关键视频帧；The selection module is configured to select a target key video frame from the plurality of candidate key video frames according to the target key index of each candidate key video frame;
    广告视频优化模块，用于获取关键视频帧前置策略，并基于所述关键视频帧前置策略和所述目标关键视频帧，对所述广告视频进行优化，得到目标广告视频。The advertising video optimization module is configured to obtain a key video frame advance strategy, and based on the key video frame advance strategy and the target key video frame, optimize the advertisement video to obtain a target advertisement video.
  9. 如权利要求8所述的广告视频优化装置，其特征在于，所述关键指数评估模块还用于：8. The advertising video optimization device of claim 8, wherein the key index evaluation module is further configured to:
    获取所述若干候选关键视频帧中每一候选关键视频帧包含的文字信息；Acquiring text information contained in each candidate key video frame among the plurality of candidate key video frames;
    依据所述关键指数评估算法和所述文字信息，确定每一候选关键视频帧的目标关键指数。According to the key index evaluation algorithm and the text information, the target key index of each candidate key video frame is determined.
  10. 如权利要求9所述的广告视频优化装置，其特征在于，所述关键指数评估模块还用于：8. The advertising video optimization device of claim 9, wherein the key index evaluation module is further configured to:
    计算所述每一候选关键视频帧包含的文字信息的熵，并计算所述每一候选关键视频帧包含的文字信息与预设文字信息的相似度；Calculating the entropy of the text information contained in each candidate key video frame, and calculating the similarity between the text information contained in each candidate key video frame and the preset text information;
    依据所述关键指数评估算法、所述文字信息的熵和所述相似度，确定每一候选关键视频帧的目标关键指数。According to the key index evaluation algorithm, the entropy of the text information and the similarity, the target key index of each candidate key video frame is determined.
  11. 如权利要求9所述的广告视频优化装置，其特征在于，所述关键指数评估模块还用于：8. The advertising video optimization device of claim 9, wherein the key index evaluation module is further configured to:
    计算所述每一候选关键视频帧包含的文字信息的熵；Calculating the entropy of the text information contained in each candidate key video frame;
    基于预设美学分数预测模型，预测所述若干候选关键视频帧中每一候选关键视频帧的美学分数；Predict the aesthetic score of each candidate key video frame among the plurality of candidate key video frames based on a preset aesthetic score prediction model;
    依据所述关键指数评估算法、所述文字信息的熵和所述美学分数，确定所述每一候选关键视频帧的目标关键指数。According to the key index evaluation algorithm, the entropy of the text information, and the aesthetic score, the target key index of each candidate key video frame is determined.
  12. 如权利要求11所述的广告视频优化装置，其特征在于，所述关键指数评估模块还用于：11. The advertising video optimization device of claim 11, wherein the key index evaluation module is further configured to:
    计算所述每一候选关键视频帧包含的文字信息与预设文字信息的相似度；Calculating the similarity between the text information contained in each candidate key video frame and the preset text information;
    依据所述关键指数评估算法、所述文字信息的熵、所述相似度和所述美学分数，确定所述每一候选关键视频帧的目标关键指数。According to the key index evaluation algorithm, the entropy of the text information, the similarity and the aesthetic score, the target key index of each candidate key video frame is determined.
  13. 如权利要求8所述的广告视频优化装置，其特征在于，所述广告视频优化装置还包括：8. The advertising video optimization device of claim 8, wherein the advertising video optimization device further comprises:
    获取模块，用于当监测到算法策略更新指令时，依据所述算法策略更新指令，获取对应的广告反馈数据、待更新的关键指数评估算法和关键视频帧前置策略；The obtaining module is used to obtain corresponding advertisement feedback data, the key index evaluation algorithm to be updated, and the key video frame advance strategy according to the algorithm strategy update instruction when the algorithm strategy update instruction is monitored;
    更新模块，用于依据所述广告反馈数据，对待更新的关键指数评估算法和关键视频帧前置策略执行更新操作。The update module is used to perform an update operation on the key index evaluation algorithm and key video frame advance strategy to be updated based on the advertisement feedback data.
  14. 如权利要求12所述的广告视频优化装置，其特征在于，所述更新模块还用于：The advertisement video optimization device according to claim 12, wherein the update module is further configured to:
    获取预设的算法策略更新模型，并将所述广告反馈数据输入至所述算法策略更新模型；Obtaining a preset algorithm strategy update model, and input the advertisement feedback data into the algorithm strategy update model;
    获取所述算法策略更新模型输出的用于更新所述关键指数评估算法的第一参数值以及用于更新所述关键视频帧前置策略的第二参数值；Acquiring a first parameter value output by the algorithm strategy update model for updating the key index evaluation algorithm and a second parameter value for updating the key video frame forward strategy;
    依据所述第一参数值更新所述关键指数评估算法，以及依据所述第二参数值更新所述关键视频帧前置策略。The key index evaluation algorithm is updated according to the first parameter value, and the key video frame advance strategy is updated according to the second parameter value.
  15. 一种广告视频优化设备，其特征在于，所述广告视频优化设备包括：存储器、处理器及存储在所述存储器上并可在所述处理器上运行的广告视频优化程序，所述广告视频优化程序被所述处理器执行时实现以下步骤：An advertisement video optimization device, characterized in that, the advertisement video optimization device includes: a memory, a processor, and an advertisement video optimization program stored in the memory and running on the processor, and the advertisement video optimization The following steps are implemented when the program is executed by the processor:
    将待优化的广告视频分割成若干广告视频片段，并通过预设的视频帧提取算法对每个广告视频片段进行关键视频帧提取，得到若干候选关键视频帧；Divide the advertisement video to be optimized into several advertisement video fragments, and extract key video frames from each advertisement video fragment through a preset video frame extraction algorithm to obtain several candidate key video frames;
    获取关键指数评估算法，并依据所述关键指数评估算法，评估所述若干候选关键视频帧中的每一候选关键视频帧，得到每一候选关键视频帧的目标关键指数；Acquiring a key index evaluation algorithm, and evaluating each candidate key video frame among the plurality of candidate key video frames according to the key index evaluation algorithm, to obtain a target key index of each candidate key video frame;
    依据所述每一候选关键视频帧的目标关键指数，从所述若干候选关键视频帧中选择目标关键视频帧；Selecting a target key video frame from the plurality of candidate key video frames according to the target key index of each candidate key video frame;
    获取关键视频帧前置策略，并基于所述关键视频帧前置策略和所述目标关键视频帧，对所述广告视频进行优化，得到目标广告视频。Acquire a key video frame advance strategy, and based on the key video frame advance strategy and the target key video frame, optimize the advertisement video to obtain a target advertisement video.
  16. 如权利要求15所述的广告视频优化设备，其特征在于，所述广告视频优化程序被所述处理器执行时，还实现以下步骤：The advertisement video optimization device according to claim 15, wherein when the advertisement video optimization program is executed by the processor, the following steps are further implemented:
    获取所述若干候选关键视频帧中每一候选关键视频帧包含的文字信息；Acquiring text information contained in each candidate key video frame among the plurality of candidate key video frames;
    依据所述关键指数评估算法和所述文字信息，确定每一候选关键视频帧的目标关键指数。According to the key index evaluation algorithm and the text information, the target key index of each candidate key video frame is determined.
  17. 如权利要求16所述的广告视频优化设备，其特征在于，所述广告视频优化程序被所述处理器执行时，还实现以下步骤：The advertisement video optimization device of claim 16, wherein when the advertisement video optimization program is executed by the processor, the following steps are further implemented:
    计算所述每一候选关键视频帧包含的文字信息的熵，并计算所述每一候选关键视频帧包含的文字信息与预设文字信息的相似度；Calculating the entropy of the text information contained in each candidate key video frame, and calculating the similarity between the text information contained in each candidate key video frame and the preset text information;
    依据所述关键指数评估算法、所述文字信息的熵和所述相似度，确定每一候选关键视频帧的目标关键指数。According to the key index evaluation algorithm, the entropy of the text information and the similarity, the target key index of each candidate key video frame is determined.
  18. 一种计算机可读存储介质，其特征在于，所述计算机可读存储介质上存储有广告视频优化程序，所述广告视频优化程序被处理器执行时实现以下步骤：A computer-readable storage medium, characterized in that an advertisement video optimization program is stored on the computer-readable storage medium, and the following steps are implemented when the advertisement video optimization program is executed by a processor:
    将待优化的广告视频分割成若干广告视频片段，并通过预设的视频帧提取算法对每个广告视频片段进行关键视频帧提取，得到若干候选关键视频帧；Divide the advertisement video to be optimized into several advertisement video fragments, and extract key video frames from each advertisement video fragment through a preset video frame extraction algorithm to obtain several candidate key video frames;
    获取关键指数评估算法，并依据所述关键指数评估算法，评估所述若干候选关键视频帧中的每一候选关键视频帧，得到每一候选关键视频帧的目标关键指数；Acquiring a key index evaluation algorithm, and evaluating each candidate key video frame among the plurality of candidate key video frames according to the key index evaluation algorithm, to obtain a target key index of each candidate key video frame;
    依据所述每一候选关键视频帧的目标关键指数，从所述若干候选关键视频帧中选择目标关键视频帧；Selecting a target key video frame from the plurality of candidate key video frames according to the target key index of each candidate key video frame;
    获取关键视频帧前置策略，并基于所述关键视频帧前置策略和所述目标关键视频帧，对所述广告视频进行优化，得到目标广告视频。Acquire a key video frame advance strategy, and based on the key video frame advance strategy and the target key video frame, optimize the advertisement video to obtain a target advertisement video.
  19. 如权利要求18所述的计算机可读存储介质，其特征在于，所述广告视频优化程序被所述处理器执行时，还实现以下步骤：18. The computer-readable storage medium of claim 18, wherein when the advertisement video optimization program is executed by the processor, the following steps are further implemented:
    获取所述若干候选关键视频帧中每一候选关键视频帧包含的文字信息；Acquiring text information contained in each candidate key video frame among the plurality of candidate key video frames;
    依据所述关键指数评估算法和所述文字信息，确定每一候选关键视频帧的目标关键指数。According to the key index evaluation algorithm and the text information, the target key index of each candidate key video frame is determined.
  20. 如权利要求19所述的计算机可读存储介质，其特征在于，所述广告视频优化程序被所述处理器执行时，还实现以下步骤：19. The computer-readable storage medium of claim 19, wherein when the advertisement video optimization program is executed by the processor, the following steps are further implemented:
    计算所述每一候选关键视频帧包含的文字信息的熵，并计算所述每一候选关键视频帧包含的文字信息与预设文字信息的相似度；Calculating the entropy of the text information contained in each candidate key video frame, and calculating the similarity between the text information contained in each candidate key video frame and the preset text information;
    依据所述关键指数评估算法、所述文字信息的熵和所述相似度，确定每一候选关键视频帧的目标关键指数。 According to the key index evaluation algorithm, the entropy of the text information and the similarity, the target key index of each candidate key video frame is determined.
PCT/CN2019/078528 2019-02-22 2019-03-18 Advertisement video optimising method, apparatus and device and computer readable storage medium WO2020168606A1 (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201910140441.4 2019-02-22
CN201910140441.4A CN109922379B (en) 2019-02-22 2019-02-22 Advertisement video optimization method, device and equipment and computer readable storage medium

Publications (1)

Publication Number Publication Date
WO2020168606A1 true WO2020168606A1 (en) 2020-08-27

Family

ID=66962344

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/078528 WO2020168606A1 (en) 2019-02-22 2019-03-18 Advertisement video optimising method, apparatus and device and computer readable storage medium

Country Status (2)

Country Link
CN (1) CN109922379B (en)
WO (1) WO2020168606A1 (en)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113569093A (en) * 2021-01-18 2021-10-29 腾讯科技（深圳）有限公司 Video processing method, video processing device, computer equipment and storage medium
CN113610557A (en) * 2021-07-09 2021-11-05 苏州众言网络科技股份有限公司 Advertisement video processing method and device
CN114697762A (en) * 2022-04-07 2022-07-01 脸萌有限公司 Processing method, processing device, terminal equipment and medium
CN114708008A (en) * 2021-12-30 2022-07-05 北京有竹居网络技术有限公司 Promotion content processing method, device, equipment, medium and product

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111104370A (en) * 2019-12-18 2020-05-05 北京大龙得天力广告传媒有限公司 Advertisement video storage system and method
CN111294646B (en) * 2020-02-17 2022-08-30 腾讯科技（深圳）有限公司 Video processing method, device, equipment and storage medium
CN113378002B (en) 2021-08-11 2022-01-21 北京达佳互联信息技术有限公司 Information delivery method and device, electronic equipment and storage medium

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20070183497A1 (en) * 2006-02-03 2007-08-09 Jiebo Luo Extracting key frame candidates from video clip
US20090094113A1 (en) * 2007-09-07 2009-04-09 Digitalsmiths Corporation Systems and Methods For Using Video Metadata to Associate Advertisements Therewith
CN107172481A (en) * 2017-05-09 2017-09-15 深圳市炜光科技有限公司 Video segment splices method of combination and system
CN107392666A (en) * 2017-07-24 2017-11-24 北京奇艺世纪科技有限公司 Advertisement data processing method, device and advertisement placement method and device

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104581380B (en) * 2014-12-30 2018-08-31 联想(北京)有限公司 A kind of method and mobile terminal of information processing
CN108632641A (en) * 2018-05-04 2018-10-09 百度在线网络技术（北京）有限公司 Method for processing video frequency and device
CN109121021A (en) * 2018-09-28 2019-01-01 北京周同科技有限公司 A kind of generation method of Video Roundup, device, electronic equipment and storage medium
CN109194978A (en) * 2018-10-15 2019-01-11 广州虎牙信息科技有限公司 Live video clipping method, device and electronic equipment

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20070183497A1 (en) * 2006-02-03 2007-08-09 Jiebo Luo Extracting key frame candidates from video clip
US20090094113A1 (en) * 2007-09-07 2009-04-09 Digitalsmiths Corporation Systems and Methods For Using Video Metadata to Associate Advertisements Therewith
CN107172481A (en) * 2017-05-09 2017-09-15 深圳市炜光科技有限公司 Video segment splices method of combination and system
CN107392666A (en) * 2017-07-24 2017-11-24 北京奇艺世纪科技有限公司 Advertisement data processing method, device and advertisement placement method and device

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113569093A (en) * 2021-01-18 2021-10-29 腾讯科技（深圳）有限公司 Video processing method, video processing device, computer equipment and storage medium
CN113610557A (en) * 2021-07-09 2021-11-05 苏州众言网络科技股份有限公司 Advertisement video processing method and device
CN114708008A (en) * 2021-12-30 2022-07-05 北京有竹居网络技术有限公司 Promotion content processing method, device, equipment, medium and product
CN114697762A (en) * 2022-04-07 2022-07-01 脸萌有限公司 Processing method, processing device, terminal equipment and medium
CN114697762B (en) * 2022-04-07 2023-11-28 脸萌有限公司 Processing method, processing device, terminal equipment and medium

Also Published As

Publication number Publication date
CN109922379A (en) 2019-06-21
CN109922379B (en) 2022-01-04

Similar Documents

Publication Publication Date Title
WO2020168606A1 (en) Advertisement video optimising method, apparatus and device and computer readable storage medium
WO2021003930A1 (en) Quality inspection method, apparatus, and device for customer service audio, and computer readable storage medium
WO2017150860A1 (en) Predicting text input based on user demographic information and context information
WO2021008017A1 (en) Federation learning method, system, terminal device and storage medium
WO2010119996A1 (en) Method and apparatus for providing moving image advertisements
WO2010024628A2 (en) Searching method using extended keyword pool and system thereof
WO2012134180A2 (en) Emotion classification method for analyzing inherent emotions in a sentence, and emotion classification method for multiple sentences using context information
EP3494498A1 (en) Electronic device and method for controlling the same
WO2018166099A1 (en) Information leakage detection method and device, server, and computer-readable storage medium
WO2017131348A1 (en) Electronic apparatus and controlling method thereof
WO2018201638A1 (en) Image recognition based information collection method, mobile terminal and storage medium
WO2015137666A1 (en) Object recognition apparatus and control method therefor
WO2016129840A1 (en) Display apparatus and information providing method thereof
WO2021029642A1 (en) System and method for recognizing user's speech
WO2020114333A1 (en) Video-based ar education method, smart television, storage medium, and device
WO2016129871A1 (en) Image processing device, and method and system for controlling image processing device
WO2015178716A1 (en) Search method and device
WO2011068315A4 (en) Apparatus for selecting optimum database using maximal concept-strength recognition technique and method thereof
WO2013032198A1 (en) Item-based recommendation engine for recommending a highly-associated item
WO2018124464A1 (en) Electronic device and search service providing method of electronic device
WO2024112108A1 (en) Real-time drm-based video streaming system and video streaming method therefor
WO2023132657A1 (en) Device, method, and program for providing product trend prediction service
WO2023058835A1 (en) Electronic device and control method therefor
WO2015126043A1 (en) Advertisement display method and advertisement providing method, and apparatus applied thereto
WO2020141745A1 (en) Method for improving reliability of artificial intelligence-based object recognition using collective intelligence-based mutual verification

Legal Events

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

Ref document number: 19916458

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: 19916458

Country of ref document: EP

Kind code of ref document: A1