CN112668575B - Key information extraction method and device, electronic equipment and storage medium - Google Patents
Key information extraction method and device, electronic equipment and storage medium Download PDFInfo
- Publication number
- CN112668575B CN112668575B CN202011581164.XA CN202011581164A CN112668575B CN 112668575 B CN112668575 B CN 112668575B CN 202011581164 A CN202011581164 A CN 202011581164A CN 112668575 B CN112668575 B CN 112668575B
- Authority
- CN
- China
- Prior art keywords
- preset
- certificate
- key information
- target
- standard
- Prior art date
- Legal status (The legal status 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 status listed.)
- Active
Links
- 238000000605 extraction Methods 0.000 title claims abstract description 44
- 238000010586 diagram Methods 0.000 claims abstract description 105
- 238000001514 detection method Methods 0.000 claims abstract description 99
- 238000012545 processing Methods 0.000 claims abstract description 96
- 238000012549 training Methods 0.000 claims abstract description 54
- 238000012937 correction Methods 0.000 claims abstract description 32
- 230000003321 amplification Effects 0.000 claims abstract description 24
- 238000003199 nucleic acid amplification method Methods 0.000 claims abstract description 24
- 238000012216 screening Methods 0.000 claims abstract description 19
- 230000009466 transformation Effects 0.000 claims description 42
- PXFBZOLANLWPMH-UHFFFAOYSA-N 16-Epiaffinine Natural products C1C(C2=CC=CC=C2N2)=C2C(=O)CC2C(=CC)CN(C)C1C2CO PXFBZOLANLWPMH-UHFFFAOYSA-N 0.000 claims description 38
- 230000006870 function Effects 0.000 claims description 19
- 238000000034 method Methods 0.000 claims description 14
- 238000013507 mapping Methods 0.000 claims description 12
- 230000004913 activation Effects 0.000 claims description 10
- 238000004364 calculation method Methods 0.000 claims description 8
- 238000002372 labelling Methods 0.000 claims description 8
- 238000013145 classification model Methods 0.000 claims description 6
- 238000004590 computer program Methods 0.000 claims description 6
- 230000001131 transforming effect Effects 0.000 claims description 5
- 230000008859 change Effects 0.000 claims description 4
- 239000011159 matrix material Substances 0.000 claims description 3
- 230000008569 process Effects 0.000 claims description 3
- 238000005516 engineering process Methods 0.000 abstract description 5
- 230000000694 effects Effects 0.000 description 8
- 238000007726 management method Methods 0.000 description 5
- 102100032202 Cornulin Human genes 0.000 description 2
- 101000920981 Homo sapiens Cornulin Proteins 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 238000006073 displacement reaction Methods 0.000 description 2
- 239000004973 liquid crystal related substance Substances 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 230000002093 peripheral effect Effects 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- 239000002023 wood Substances 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 238000013434 data augmentation Methods 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000000802 evaporation-induced self-assembly Methods 0.000 description 1
- 239000012634 fragment Substances 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Landscapes
- Character Input (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to an intelligent decision making technology, and discloses a key information extraction method, which comprises the following steps: performing data amplification processing on the original document image to obtain an initial training image; training a preset lightweight detection model by using the initial training image to obtain a trained lightweight detection model; obtaining a certificate image to be identified, and carrying out target area screening treatment on the certificate image to be identified to obtain a certificate area diagram; performing angle correction processing on the certificate regional graph to obtain a standard regional graph; performing target detection processing on the standard area diagram to obtain a target rectangular diagram; and carrying out text recognition processing on the target rectangular chart to obtain key information. In addition, the invention also relates to a blockchain technology, and the certificate area graph can be stored in nodes of the blockchain. The invention also provides a key information extraction device, electronic equipment and a computer readable storage medium. The invention can solve the problem of low accuracy of extracting the key information.
Description
Technical Field
The present invention relates to the field of intelligent decision making technologies, and in particular, to a method and apparatus for extracting key information, an electronic device, and a computer readable storage medium.
Background
With the development of information technology, more and more online services (for example, online business handling of hydropower business and online business handling of banking business) are performed, and verification of user identity is generally required during online business handling. When identity verification is performed, it is generally necessary to identify a user's certificate, and then determine the identity of the user according to the content of the certificate. The identification of a user's credentials typically requires the extraction of critical information from the credentials, such as: the identity of the user is confirmed by checking the key information.
The existing key information extraction method is to match the cut certificate picture with the existing template to obtain the key information fragment of the certificate. The method is easily affected by the template, and the accuracy of extracting the key information is not high.
Disclosure of Invention
The invention provides a key information extraction method, a key information extraction device and a computer readable storage medium, and mainly aims to solve the problem of low accuracy of key information extraction.
In order to achieve the above object, the present invention provides a key information extraction method, including:
Acquiring an original document image, and performing data amplification processing on the original document image to obtain an initial training image;
Training a preset lightweight detection model by using the initial training image to obtain a trained lightweight detection model;
Obtaining a certificate image to be identified, and performing target area screening treatment on the certificate image to be identified by using the trained lightweight detection model to obtain a certificate area diagram;
performing angle correction processing on the certificate regional graph to obtain a standard regional graph;
performing target detection processing on the standard region map by using a preset target detection model in any direction to obtain a target rectangular map;
And carrying out text recognition processing on the target rectangular graph to obtain key information.
Optionally, the performing angle correction processing on the certificate area map to obtain a standard area map includes:
Performing angle prediction processing on the certificate area diagram by using a preset four-classification model to obtain angle information of the certificate area diagram;
Judging whether the angle information accords with a preset angle standard or not;
If the angle information accords with the preset angle standard, judging that the certificate area diagram is a standard area diagram;
And if the angle information does not accord with the preset angle standard, carrying out angle correction on the certificate area diagram to obtain a standard area diagram.
Optionally, the performing target detection processing on the standard area map by using a preset target detection model in any direction to obtain a target rectangular map includes:
Carrying out rectangular frame labeling on the relevant areas of the key information in the standard area diagram by using a preset target detection model in any direction;
Acquiring a twiddle factor corresponding to the marked rectangular frame;
judging whether the twiddle factor is a twiddle factor threshold;
If the twiddle factor is a twiddle factor threshold, capturing a picture corresponding to the rectangular frame from the standard region picture to obtain a target rectangular picture;
and if the twiddle factor is not the twiddle factor threshold, carrying out affine transformation on the rectangular frame, and intercepting pictures corresponding to the rectangular frame after affine transformation from the standard region diagram to obtain a target rectangular diagram.
Optionally, the affine transforming the rectangular frame includes:
Mapping the rectangular frame to a preset two-dimensional rectangular coordinate system, and extracting coordinate values of the rectangular frame;
Transforming the coordinate values according to a preset affine transformation formula to obtain transformed coordinate values;
Mapping the transformation coordinate values on the two-dimensional rectangular coordinate system to obtain a rectangular frame after affine transformation.
Optionally, the text recognition processing is performed on the target rectangle graph to obtain key information, including:
performing feature extraction processing on the target rectangular graph to obtain a feature sequence, wherein the feature sequence comprises a plurality of components;
probability calculation is carried out on the components by using a preset activation function, so that probability values of the components are obtained;
and determining the component corresponding to the maximum probability value as key information.
Optionally, training the preset lightweight detection model by using the initial training image to obtain a trained lightweight detection model, including:
Performing frame selection processing on the initial training image by using a preset priori frame to obtain a prediction area diagram;
calculating the coincidence value between the predicted area diagram and a preset real area diagram according to a coincidence value formula;
And when the coincidence value is smaller than a preset threshold value, adjusting the internal parameters of the preset lightweight detection model until the coincidence value is larger than or equal to the preset threshold value, and obtaining the trained lightweight detection model.
Optionally, the coincidence value formula includes:
wherein, IOU is the coincidence value, detectionResult is the predicted region map, groundTruth is the real region map, detectionResult n GroundTruth is the intersection between the predicted region map and the real region map, detectionResult n GroundTruth is the union between the predicted region map and the real region map.
In order to solve the above problems, the present invention also provides a key information extraction apparatus, the apparatus comprising:
the data amplification module is used for acquiring an original document image, and carrying out data amplification processing on the original document image to obtain an initial training image;
The model training module is used for training a preset lightweight detection model by utilizing the initial training image to obtain a trained lightweight detection model;
the regional screening module is used for acquiring a certificate image to be identified, and performing target regional screening treatment on the certificate image to be identified by utilizing the trained lightweight detection model to obtain a certificate regional graph;
The angle correction module is used for performing angle correction processing on the certificate area map to obtain a standard area map;
The target detection module is used for carrying out target detection processing on the standard region graph by utilizing a preset target detection model in any direction to obtain a target rectangular graph;
and the text recognition module is used for carrying out text recognition processing on the target rectangle graph to obtain key information.
In order to solve the above-mentioned problems, the present invention also provides an electronic apparatus including:
A memory storing at least one instruction; and
And the processor executes the instructions stored in the memory to realize the key information extraction method.
In order to solve the above-mentioned problems, the present invention also provides a computer-readable storage medium having stored therein at least one instruction that is executed by a processor in an electronic device to implement the above-mentioned key information extraction method.
According to the invention, the original training image is obtained by carrying out data amplification processing on the original certificate image, the data amplification processing can enlarge the quantity of model training data, and the robustness and the accuracy of a trained lightweight detection model are increased; after the certificate image to be identified is obtained, the trained light detection model can be used for accurately and rapidly obtaining the certificate area diagram, and then the angle correction processing is carried out on the certificate area diagram, so that the target rectangular frame can be conveniently and rapidly obtained, the text recognition processing is carried out on the rectangular frame, key information is obtained, the accuracy of extracting the key information is improved, and meanwhile, the efficiency of extracting the key information is also improved. Therefore, the key information extraction method, the device, the electronic equipment and the computer readable storage medium can solve the problem of low accuracy of key information extraction.
Drawings
Fig. 1 is a flow chart of a key information extraction method according to an embodiment of the invention;
FIG. 2 is a functional block diagram of a key information extraction device according to an embodiment of the present invention;
fig. 3 is a schematic structural diagram of an electronic device for implementing the key information extraction method according to an embodiment of the present invention.
The achievement of the objects, functional features and advantages of the present invention will be further described with reference to the accompanying drawings, in conjunction with the embodiments.
Detailed Description
It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.
The embodiment of the application provides a key information extraction method. The execution subject of the key information extraction method includes, but is not limited to, at least one of a server, a terminal, and the like, which can be configured to execute the method provided by the embodiment of the application. In other words, the key information extraction method may be performed by software or hardware installed in a terminal device or a server device, and the software may be a blockchain platform. The service end includes but is not limited to: a single server, a server cluster, a cloud server or a cloud server cluster, and the like.
Referring to fig. 1, a flow chart of a key information extraction method according to an embodiment of the invention is shown. In this embodiment, the key information extraction method includes:
S1, acquiring an original document image, and performing data amplification processing on the original document image to obtain an initial training image.
In the embodiment of the invention, the certificate image is an image of a certificate shot by a camera. For example, images of credentials taken by a camera on a mobile electronic device (e.g., a cell phone).
Specifically, document images include, but are not limited to: an image of an identity card, an image of a social security card, an image of a passport.
The data amplification processing comprises random color dithering, random brightness dithering, random saturation dithering and random contrast dithering.
Specifically, the random color dithering is to generate displacement to the hue of the formed image, so as to cause the color crossing effect of adjacent point-like differences; the random brightness dithering is an effect of causing bright-dark cross on an image; the random saturation dithering is used for generating a saturation difference-like cross effect; the random contrast dithering is a crossover effect that creates contrast differences.
In detail, in the embodiment of the invention, the data amplification processing is performed on the original document image, so that the quantity of model training data can be increased, and the robustness of the model is improved.
And S2, training a preset lightweight detection model by using the initial training image to obtain a trained lightweight detection model.
In the embodiment of the present invention, training a preset lightweight detection model by using the initial training image to obtain a trained lightweight detection model includes:
Performing frame selection processing on the initial training image by using a preset priori frame to obtain a prediction area diagram;
calculating the coincidence value between the predicted area diagram and a preset real area diagram according to a coincidence value formula;
And when the coincidence value is smaller than a preset threshold value, adjusting the internal parameters of the preset lightweight detection model until the coincidence value is larger than or equal to the preset threshold value, and obtaining the trained lightweight detection model.
In detail, the preset prior frame is placed on the initial training image for frame selection, the picture selected by the frame is the predicted area picture, a preset coincidence value formula is utilized to calculate a coincidence value between the predicted area picture and the real area picture, the coincidence value is used for judging the similarity degree between the predicted area picture and the real area picture, when the coincidence value is smaller than a preset threshold value, the similarity degree between the predicted area picture and the real area picture does not reach a preset standard, and the internal parameters of a preset lightweight detection model are required to be adjusted, wherein the internal parameters can be model weights or model gradients until the coincidence value is larger than or equal to the preset threshold value, and the trained lightweight detection model is obtained.
Wherein the preset threshold value may be 0.5.
Specifically, the coincidence value formula includes:
wherein, IOU is the coincidence value, detectionResult is the predicted region map, groundTruth is the real region map, detectionResult n GroundTruth is the intersection between the predicted region map and the real region map, detectionResult n GroundTruth is the union between the predicted region map and the real region map.
In detail, an intersection between the predicted area map and the real area map and a union between the predicted area map and the real area map are respectively obtained, and a preset coincidence value formula is combined to calculate a coincidence value, wherein in the embodiment of the invention, the predicted area map is a predicted certificate area picture, and the real area map is an existing standard certificate area picture.
S3, acquiring a certificate image to be identified, and performing target area screening treatment on the certificate image to be identified by using the trained lightweight detection model to obtain a certificate area diagram.
The target area refers to an area of certificate information in the certificate image to be identified.
For example, the document image to be identified includes a background image (such as wood grain background) and a document image, and then the trained lightweight detection model is used to perform target area screening processing on the document image to be identified, so as to obtain the document image therein, namely the document area map.
In the embodiment of the invention, the target area is intercepted from the certificate image to be identified by using the trained lightweight detection model and is used as the certificate area diagram, the irrelevant information interference of the background in the certificate image to be identified is removed, the duty ratio of the information in the certificate area diagram in the certificate image to be identified is increased, the information redundancy is reduced, and the identification efficiency is improved.
And S4, performing angle correction processing on the certificate regional graph to obtain a standard regional graph.
In the embodiment of the present invention, the performing an angle correction process on the certificate area map to obtain a standard area map includes:
Performing angle prediction processing on the certificate area diagram by using a preset four-classification model to obtain angle information of the certificate area diagram;
Judging whether the angle information accords with a preset angle standard or not;
If the angle information accords with the preset angle standard, judging that the certificate area diagram is a standard area diagram;
And if the angle information does not accord with the preset angle standard, carrying out angle correction on the certificate area diagram to obtain a standard area diagram.
In detail, angle prediction processing is performed on the certificate area diagram by using a preset four-classification model, so that angle information of the certificate area diagram is obtained, the angle information is a deflection angle of the certificate area diagram on the original certificate image, the placement direction of the certificate area diagram can be judged according to the angle information, different placement directions can influence a subsequent certificate recognition result, whether the angle information accords with a preset angle standard or not is judged, the preset angle standard is 0 degree, certificate recognition is facilitated, and if the angle information accords with the preset angle standard, the certificate area diagram is judged to be a standard area diagram;
and if the angle information does not accord with the preset angle standard, carrying out angle correction on the certificate area diagram, and obtaining a standard area diagram by correcting the angle of the certificate area diagram to be 0 degrees.
Specifically, in the embodiment of the present invention, the angle information includes a deviation angle of the document area map on the original document image, for example, 0 degrees, 90 degrees, or 180 degrees.
The preset angle standard means that the deflection angle of the certificate area diagram on the original certificate image is 0 degrees.
Further, the angular correction of the document area map according to the angular information is to rotationally correct the document area map to conform to an angular standard, for example, to rotationally correct the document area map to a direction approaching 0 degrees.
S5, performing target detection processing on the standard region map by using a preset target detection model in any direction to obtain a target rectangular map.
In the embodiment of the present invention, the target detection processing is performed on the standard area map by using a preset target detection model in any direction to obtain a target rectangular map, including:
Carrying out rectangular frame labeling on the relevant areas of the key information in the standard area diagram by using a preset target detection model in any direction;
Acquiring a twiddle factor corresponding to the marked rectangular frame;
judging whether the twiddle factor is a twiddle factor threshold;
If the twiddle factor is a twiddle factor threshold, capturing a picture corresponding to the rectangular frame from the standard region picture to obtain a target rectangular picture;
and if the twiddle factor is not the twiddle factor threshold, carrying out affine transformation on the rectangular frame, and intercepting pictures corresponding to the rectangular frame after affine transformation from the standard region diagram to obtain a target rectangular diagram.
The target detection model in any direction can be a model with a Faster-rcnn structure.
In detail, rectangular labeling is performed on a relevant area of key information in the standard area diagram according to a preset target detection model in any direction, for example, the standard area diagram is a processed identity card picture, various key information including but not limited to name, gender, identity card number, home address and the like exists on the identity card picture, rectangular frame labeling is performed on the key information, a twiddle factor corresponding to the rectangular frame is obtained, whether the twiddle factor is a twiddle factor threshold is judged, the twiddle factor threshold is a standard for judging whether the rectangular frame is horizontal, if the twiddle factor is a twiddle factor threshold, a picture corresponding to the rectangular frame is taken from the standard area diagram, a target rectangular diagram is obtained, if the twiddle factor is not a twiddle factor threshold, affine transformation is performed on the rectangular frame, and a picture corresponding to the rectangular frame after affine transformation is taken from the standard area diagram, and the target rectangular diagram is obtained.
Further, the affine transformation of the rectangular frame includes:
Mapping the rectangular frame to a preset two-dimensional rectangular coordinate system, and extracting coordinate values of the rectangular frame;
Transforming the coordinate values according to a preset affine transformation formula to obtain transformed coordinate values;
Mapping the transformation coordinate values on the two-dimensional rectangular coordinate system to obtain a rectangular frame after affine transformation.
In detail, carrying out affine transformation on the rectangular frame mainly according to coordinate values of the rectangular frame to obtain a preset two-dimensional rectangular coordinate system, mapping the rectangular frame onto the two-dimensional rectangular coordinate system to obtain the coordinate values of the rectangular frame, carrying out transformation processing on the coordinate values according to a preset affine transformation formula with corresponding affine transformation formulas to obtain transformed coordinate values; mapping the transformation coordinate values on the two-dimensional rectangular coordinate system to obtain a rectangular frame after affine transformation.
Wherein the preset affine transformation formula is as followsWherein (x ', y') is a transformed coordinate value, (x, y) is a coordinate value,/>Is a preset affine change matrix.
And S6, performing text recognition processing on the target rectangular chart to obtain key information.
When a plurality of target rectangular charts are obtained through the key information extraction method, text recognition processing is carried out on the plurality of target rectangular charts, and a plurality of key information can be obtained. For example, a name and an identification number, or a name, an identification number, and a residence address.
In the embodiment of the invention, the text recognition processing is carried out on the target rectangular chart by using a preset text recognition model, so that key information is obtained.
In the embodiment of the present invention, the text recognition model may be a CRNN model.
Specifically, the text recognition processing is performed on the target rectangle graph to obtain key information, which includes:
performing feature extraction processing on the target rectangular graph to obtain a feature sequence, wherein the feature sequence comprises a plurality of components;
probability calculation is carried out on the components by using a preset activation function, so that probability values of the components are obtained;
and determining the component corresponding to the maximum probability value as key information.
In detail, in the embodiment of the invention, feature extraction processing is performed on the target rectangular chart to obtain a feature sequence, namely, the feature sequence in the certificate picture subjected to screening, angle correction and target detection processing is extracted, and probability calculation is performed on the components by using a preset activation function to obtain probability values of the components; the preset activation function may be a softmax function, and the component corresponding to the maximum probability value is determined to be key information.
Specifically, the convolution layer in the text recognition model may be used to convert the target rectangular graph into a feature graph with feature information, so as to obtain a feature sequence, where the feature sequence includes multiple components, probability calculation is performed on the components by using a preset activation function, probability values of the components are obtained, and the component with the largest corresponding probability in the components is used as key information.
Further, in an embodiment of the present invention, after the obtaining the key information, the method further includes: and carrying out identity recognition on the key information.
Specifically, the identifying the key information may include: and matching the key information with data in the identity information library to determine the identity of the user corresponding to the certificate image to be identified, and further determining whether to open a certain authority for the user.
According to the invention, the original training image is obtained by carrying out data amplification processing on the original certificate image, the data amplification processing can enlarge the quantity of model training data, and the robustness and the accuracy of a trained lightweight detection model are increased; after the certificate image to be identified is obtained, the trained light detection model can be used for accurately and rapidly obtaining the certificate area diagram, and then the angle correction processing is carried out on the certificate area diagram, so that the target rectangular frame can be conveniently and rapidly obtained, the text recognition processing is carried out on the rectangular frame, key information is obtained, the accuracy of extracting the key information is improved, and meanwhile, the efficiency of extracting the key information is also improved. Therefore, the key information extraction method provided by the invention can solve the problem of low accuracy of extracting the key information.
Fig. 2 is a functional block diagram of a key information extraction device according to an embodiment of the present invention.
The key information extraction apparatus 100 of the present invention may be installed in an electronic device. Depending on the functions implemented, the key information extraction device 100 may include a data augmentation module 101, a model training module 102, a region screening module 103, an angle correction module 104, a target detection module 105, and a text recognition module 106. The module of the invention, which may also be referred to as a unit, refers to a series of computer program segments, which are stored in the memory of the electronic device, capable of being executed by the processor of the electronic device and of performing a fixed function.
In the present embodiment, the functions concerning the respective modules/units are as follows:
the data amplification module 101 is configured to obtain an original document image, and perform data amplification processing on the original document image to obtain an initial training image;
the model training module 102 is configured to train a preset lightweight detection model by using the initial training image to obtain a trained lightweight detection model;
The region screening module 103 is configured to obtain a document image to be identified, and perform target region screening processing on the document image to be identified by using the trained lightweight detection model to obtain a document region map;
the angle correction module 104 is configured to perform angle correction processing on the certificate area map to obtain a standard area map;
the target detection module 105 is configured to perform target detection processing on the standard area map by using a preset target detection model in any direction, so as to obtain a target rectangular map;
The text recognition module 106 is configured to perform text recognition processing on the target rectangle chart to obtain key information.
The data amplification module 101 is configured to obtain an original document image, and perform data amplification processing on the original document image to obtain an initial training image.
In the embodiment of the invention, the certificate image is an image of a certificate shot by a camera. For example, images of credentials taken by a camera on a mobile electronic device (e.g., a cell phone).
Specifically, document images include, but are not limited to: an image of an identity card, an image of a social security card, an image of a passport.
The data amplification processing comprises random color dithering, random brightness dithering, random saturation dithering and random contrast dithering.
Specifically, the random color dithering is to generate displacement to the hue of the formed image, so as to cause the color crossing effect of adjacent point-like differences; the random brightness dithering is an effect of causing bright-dark cross on an image; the random saturation dithering is used for generating a saturation difference-like cross effect; the random contrast dithering is a crossover effect that creates contrast differences.
In detail, in the embodiment of the invention, the data amplification processing is performed on the original document image, so that the quantity of model training data can be increased, and the robustness of the model is improved.
The model training module 102 is configured to train a preset lightweight detection model by using the initial training image, so as to obtain a trained lightweight detection model.
In the embodiment of the present invention, the model training module 102 is specifically configured to:
Performing frame selection processing on the initial training image by using a preset priori frame to obtain a prediction area diagram;
calculating the coincidence value between the predicted area diagram and a preset real area diagram according to a coincidence value formula;
And when the coincidence value is smaller than a preset threshold value, adjusting the internal parameters of the preset lightweight detection model until the coincidence value is larger than or equal to the preset threshold value, and obtaining the trained lightweight detection model.
In detail, the preset prior frame is placed on the initial training image for frame selection, the picture selected by the frame is the predicted area picture, a preset coincidence value formula is utilized to calculate a coincidence value between the predicted area picture and the real area picture, the coincidence value is used for judging the similarity degree between the predicted area picture and the real area picture, when the coincidence value is smaller than a preset threshold value, the similarity degree between the predicted area picture and the real area picture does not reach a preset standard, and the internal parameters of a preset lightweight detection model are required to be adjusted, wherein the internal parameters can be model weights or model gradients until the coincidence value is larger than or equal to the preset threshold value, and the trained lightweight detection model is obtained.
Wherein the preset threshold value may be 0.5.
Specifically, the coincidence value formula includes:
wherein, IOU is the coincidence value, detectionResult is the predicted region map, groundTruth is the real region map, detectionResult n GroundTruth is the intersection between the predicted region map and the real region map, detectionResult n GroundTruth is the union between the predicted region map and the real region map.
In detail, an intersection between the predicted area map and the real area map and a union between the predicted area map and the real area map are respectively obtained, and a preset coincidence value formula is combined to calculate a coincidence value, wherein in the embodiment of the invention, the predicted area map is a predicted certificate area picture, and the real area map is an existing standard certificate area picture.
The region screening module 103 is configured to obtain a document image to be identified, and perform target region screening processing on the document image to be identified by using the trained lightweight detection model, so as to obtain a document region map.
The target area refers to an area of certificate information in the certificate image to be identified.
For example, the document image to be identified includes a background image (such as wood grain background) and a document image, and then the trained lightweight detection model is used to perform target area screening processing on the document image to be identified, so as to obtain the document image therein, namely the document area map.
In the embodiment of the invention, the target area is intercepted from the certificate image to be identified by using the trained lightweight detection model and is used as the certificate area diagram, the irrelevant information interference of the background in the certificate image to be identified is removed, the duty ratio of the information in the certificate area diagram in the certificate image to be identified is increased, the information redundancy is reduced, and the identification efficiency is improved.
The angle correction module 104 is configured to perform angle correction processing on the certificate area map to obtain a standard area map.
In the embodiment of the present invention, the angle correction module 104 is specifically configured to:
Performing angle prediction processing on the certificate area diagram by using a preset four-classification model to obtain angle information of the certificate area diagram;
Judging whether the angle information accords with a preset angle standard or not;
If the angle information accords with the preset angle standard, judging that the certificate area diagram is a standard area diagram;
And if the angle information does not accord with the preset angle standard, carrying out angle correction on the certificate area diagram to obtain a standard area diagram.
In detail, angle prediction processing is performed on the certificate area diagram by using a preset four-classification model, so that angle information of the certificate area diagram is obtained, the angle information is a deflection angle of the certificate area diagram on the original certificate image, the placement direction of the certificate area diagram can be judged according to the angle information, different placement directions can influence a subsequent certificate recognition result, whether the angle information accords with a preset angle standard or not is judged, the preset angle standard is 0 degree, certificate recognition is facilitated, and if the angle information accords with the preset angle standard, the certificate area diagram is judged to be a standard area diagram;
and if the angle information does not accord with the preset angle standard, carrying out angle correction on the certificate area diagram, and obtaining a standard area diagram by correcting the angle of the certificate area diagram to be 0 degrees.
Specifically, in the embodiment of the present invention, the angle information includes a deviation angle of the document area map on the original document image, for example, 0 degrees, 90 degrees, or 180 degrees.
The preset angle standard means that the deflection angle of the certificate area diagram on the original certificate image is 0 degrees.
Further, the angular correction of the document area map according to the angular information is to rotationally correct the document area map to conform to an angular standard, for example, to rotationally correct the document area map to a direction approaching 0 degrees.
The target detection module 105 is configured to perform target detection processing on the standard area map by using a preset target detection model in any direction, so as to obtain a target rectangular map.
In the embodiment of the present invention, the target detection module 105 is specifically configured to:
Carrying out rectangular frame labeling on the relevant areas of the key information in the standard area diagram by using a preset target detection model in any direction;
Acquiring a twiddle factor corresponding to the marked rectangular frame;
judging whether the twiddle factor is a twiddle factor threshold;
If the twiddle factor is a twiddle factor threshold, capturing a picture corresponding to the rectangular frame from the standard region picture to obtain a target rectangular picture;
and if the twiddle factor is not the twiddle factor threshold, carrying out affine transformation on the rectangular frame, and intercepting pictures corresponding to the rectangular frame after affine transformation from the standard region diagram to obtain a target rectangular diagram.
The target detection model in any direction can be a model with a Faster-rcnn structure.
In detail, rectangular labeling is performed on a relevant area of key information in the standard area diagram according to a preset target detection model in any direction, for example, the standard area diagram is a processed identity card picture, various key information including but not limited to name, gender, identity card number, home address and the like exists on the identity card picture, rectangular frame labeling is performed on the key information, a twiddle factor corresponding to the rectangular frame is obtained, whether the twiddle factor is a twiddle factor threshold is judged, the twiddle factor threshold is a standard for judging whether the rectangular frame is horizontal, if the twiddle factor is a twiddle factor threshold, a picture corresponding to the rectangular frame is taken from the standard area diagram, a target rectangular diagram is obtained, if the twiddle factor is not a twiddle factor threshold, affine transformation is performed on the rectangular frame, and a picture corresponding to the rectangular frame after affine transformation is taken from the standard area diagram, and the target rectangular diagram is obtained.
Further, the affine transformation of the rectangular frame includes:
Mapping the rectangular frame to a preset two-dimensional rectangular coordinate system, and extracting coordinate values of the rectangular frame;
Transforming the coordinate values according to a preset affine transformation formula to obtain transformed coordinate values;
Mapping the transformation coordinate values on the two-dimensional rectangular coordinate system to obtain a rectangular frame after affine transformation.
In detail, carrying out affine transformation on the rectangular frame mainly according to coordinate values of the rectangular frame to obtain a preset two-dimensional rectangular coordinate system, mapping the rectangular frame onto the two-dimensional rectangular coordinate system to obtain the coordinate values of the rectangular frame, carrying out transformation processing on the coordinate values according to a preset affine transformation formula with corresponding affine transformation formulas to obtain transformed coordinate values; mapping the transformation coordinate values on the two-dimensional rectangular coordinate system to obtain a rectangular frame after affine transformation.
Wherein the preset affine transformation formula is as followsWherein (x ', y') is a transformed coordinate value, (x, y) is a coordinate value,/>Is a preset affine change matrix.
The text recognition module 106 is configured to perform text recognition processing on the target rectangle chart to obtain key information.
When a plurality of target rectangular charts are obtained through the key information extraction method, text recognition processing is carried out on the plurality of target rectangular charts, and a plurality of key information can be obtained. For example, a name and an identification number, or a name, an identification number, and a residence address.
In the embodiment of the invention, the text recognition processing is carried out on the target rectangular chart by using a preset text recognition model, so that key information is obtained.
In the embodiment of the present invention, the text recognition model may be a CRNN model.
Specifically, the text recognition processing is performed on the target rectangle graph to obtain key information, which includes:
performing feature extraction processing on the target rectangular graph to obtain a feature sequence, wherein the feature sequence comprises a plurality of components;
probability calculation is carried out on the components by using a preset activation function, so that probability values of the components are obtained;
and determining the component corresponding to the maximum probability value as key information.
In detail, in the embodiment of the invention, feature extraction processing is performed on the target rectangular chart to obtain a feature sequence, namely, the feature sequence in the certificate picture subjected to screening, angle correction and target detection processing is extracted, and probability calculation is performed on the components by using a preset activation function to obtain probability values of the components; the preset activation function may be a softmax function, and the component corresponding to the maximum probability value is determined to be key information.
Specifically, the convolution layer in the text recognition model may be used to convert the target rectangular graph into a feature graph with feature information, so as to obtain a feature sequence, where the feature sequence includes multiple components, probability calculation is performed on the components by using a preset activation function, probability values of the components are obtained, and the component with the largest corresponding probability in the components is used as key information.
Further, in an embodiment of the present invention, the apparatus further includes an identity module, where the identity module is configured to: and after obtaining the key information, carrying out identity recognition on the key information.
Specifically, the identifying the key information may include: and matching the key information with data in the identity information library to determine the identity of the user corresponding to the certificate image to be identified, and further determining whether to open a certain authority for the user.
According to the invention, the original training image is obtained by carrying out data amplification processing on the original certificate image, the data amplification processing can enlarge the quantity of model training data, and the robustness and the accuracy of a trained lightweight detection model are increased; after the certificate image to be identified is obtained, the trained light detection model can be used for accurately and rapidly obtaining the certificate area diagram, and then the angle correction processing is carried out on the certificate area diagram, so that the target rectangular frame can be conveniently and rapidly obtained, the text recognition processing is carried out on the rectangular frame, key information is obtained, the accuracy of extracting the key information is improved, and meanwhile, the efficiency of extracting the key information is also improved. Therefore, the key information extraction device provided by the invention can solve the problem of low accuracy of extracting the key information.
Fig. 3 is a schematic structural diagram of an electronic device for implementing a key information extraction method according to an embodiment of the present invention.
The electronic device 1 may comprise a processor 10, a memory 11 and a bus, and may further comprise a computer program, such as a key information extraction program 12, stored in the memory 11 and executable on the processor 10.
The memory 11 includes at least one type of readable storage medium, including flash memory, a mobile hard disk, a multimedia card, a card memory (e.g., SD or DX memory, etc.), a magnetic memory, a magnetic disk, an optical disk, etc. The memory 11 may in some embodiments be an internal storage unit of the electronic device 1, such as a removable hard disk of the electronic device 1. The memory 11 may in other embodiments also be an external storage device of the electronic device 1, such as a plug-in mobile hard disk, a smart memory card (SMART MEDIA CARD, SMC), a Secure Digital (SD) card, a flash memory card (FLASH CARD) or the like, which are provided on the electronic device 1. Further, the memory 11 may also include both an internal storage unit and an external storage device of the electronic device 1. The memory 11 may be used not only for storing application software installed in the electronic device 1 and various types of data, such as codes of the key information extraction program 12, but also for temporarily storing data that has been output or is to be output.
The processor 10 may be comprised of integrated circuits in some embodiments, for example, a single packaged integrated circuit, or may be comprised of multiple integrated circuits packaged with the same or different functions, including one or more central processing units (Central Processing unit, CPU), microprocessors, digital processing chips, graphics processors, combinations of various control chips, and the like. The processor 10 is a Control Unit (Control Unit) of the electronic device, connects respective components of the entire electronic device using various interfaces and lines, and executes various functions of the electronic device 1 and processes data by running or executing programs or modules (e.g., key information extraction programs, etc.) stored in the memory 11, and calling data stored in the memory 11.
The bus may be a peripheral component interconnect standard (PERIPHERAL COMPONENT INTERCONNECT, PCI) bus, or an extended industry standard architecture (extended industry standard architecture, EISA) bus, among others. The bus may be classified as an address bus, a data bus, a control bus, etc. The bus is arranged to enable a connection communication between the memory 11 and at least one processor 10 etc.
Fig. 3 shows only an electronic device with components, it being understood by a person skilled in the art that the structure shown in fig. 3 does not constitute a limitation of the electronic device 1, and may comprise fewer or more components than shown, or may combine certain components, or may be arranged in different components.
For example, although not shown, the electronic device 1 may further include a power source (such as a battery) for supplying power to each component, and preferably, the power source may be logically connected to the at least one processor 10 through a power management device, so that functions of charge management, discharge management, power consumption management, and the like are implemented through the power management device. The power supply may also include one or more of any of a direct current or alternating current power supply, recharging device, power failure detection circuit, power converter or inverter, power status indicator, etc. The electronic device 1 may further include various sensors, bluetooth modules, wi-Fi modules, etc., which will not be described herein.
Further, the electronic device 1 may also comprise a network interface, optionally the network interface may comprise a wired interface and/or a wireless interface (e.g. WI-FI interface, bluetooth interface, etc.), typically used for establishing a communication connection between the electronic device 1 and other electronic devices.
The electronic device 1 may optionally further comprise a user interface, which may be a Display, an input unit, such as a Keyboard (Keyboard), or a standard wired interface, a wireless interface. Alternatively, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode) touch, or the like. The display may also be referred to as a display screen or display unit, as appropriate, for displaying information processed in the electronic device 1 and for displaying a visual user interface.
It should be understood that the embodiments described are for illustrative purposes only and are not limited to this configuration in the scope of the patent application.
The key information extraction program 12 stored in the memory 11 in the electronic device 1 is a combination of a plurality of instructions, which when executed in the processor 10, can implement:
Acquiring an original document image, and performing data amplification processing on the original document image to obtain an initial training image;
Training a preset lightweight detection model by using the initial training image to obtain a trained lightweight detection model;
Obtaining a certificate image to be identified, and performing target area screening treatment on the certificate image to be identified by using the trained lightweight detection model to obtain a certificate area diagram;
performing angle correction processing on the certificate regional graph to obtain a standard regional graph;
performing target detection processing on the standard region map by using a preset target detection model in any direction to obtain a target rectangular map;
And carrying out text recognition processing on the target rectangular graph to obtain key information.
Specifically, the specific implementation method of the above instructions by the processor 10 may refer to descriptions of related steps in the corresponding embodiments of fig. 1 to 3, which are not repeated herein.
Further, the modules/units integrated in the electronic device 1 may be stored in a computer readable storage medium if implemented in the form of software functional units and sold or used as separate products. The computer readable storage medium may be volatile or nonvolatile. For example, the computer readable medium may include: any entity or device capable of carrying the computer program code, a recording medium, a U disk, a removable hard disk, a magnetic disk, an optical disk, a computer Memory, a Read-Only Memory (ROM).
The present invention also provides a computer readable storage medium storing a computer program which, when executed by a processor of an electronic device, can implement:
Acquiring an original document image, and performing data amplification processing on the original document image to obtain an initial training image;
Training a preset lightweight detection model by using the initial training image to obtain a trained lightweight detection model;
Obtaining a certificate image to be identified, and performing target area screening treatment on the certificate image to be identified by using the trained lightweight detection model to obtain a certificate area diagram;
performing angle correction processing on the certificate regional graph to obtain a standard regional graph;
performing target detection processing on the standard region map by using a preset target detection model in any direction to obtain a target rectangular map;
And carrying out text recognition processing on the target rectangular graph to obtain key information.
In the several embodiments provided in the present invention, it should be understood that the disclosed apparatus, device and method may be implemented in other manners. For example, the above-described apparatus embodiments are merely illustrative, and for example, the division of the modules is merely a logical function division, and there may be other manners of division when actually implemented.
The modules described as separate components may or may not be physically separate, and components shown as modules may or may not be physical units, may be located in one place, or may be distributed over multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
In addition, each functional module in the embodiments of the present invention may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit. The integrated units can be realized in a form of hardware or a form of hardware and a form of software functional modules.
It will be evident to those skilled in the art that the invention is not limited to the details of the foregoing illustrative embodiments, and that the present invention may be embodied in other specific forms without departing from the spirit or essential characteristics thereof.
The present embodiments are, therefore, to be considered in all respects as illustrative and not restrictive, the scope of the invention being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein. Any reference signs in the claims shall not be construed as limiting the claim concerned.
The blockchain is a novel application mode of computer technologies such as distributed data storage, point-to-point transmission, consensus mechanism, encryption algorithm and the like. The blockchain (Blockchain), essentially a de-centralized database, is a string of data blocks that are generated in association using cryptographic methods, each of which contains information from a batch of network transactions for verifying the validity (anti-counterfeit) of its information and generating the next block. The blockchain may include a blockchain underlying platform, a platform product services layer, an application services layer, and the like.
Furthermore, it is evident that the word "comprising" does not exclude other elements or steps, and that the singular does not exclude a plurality. A plurality of units or means recited in the system claims can also be implemented by means of software or hardware by means of one unit or means. The terms second, etc. are used to denote a name, but not any particular order.
Finally, it should be noted that the above-mentioned embodiments are merely for illustrating the technical solution of the present invention and not for limiting the same, and although the present invention has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that modifications and equivalents may be made to the technical solution of the present invention without departing from the spirit and scope of the technical solution of the present invention.
Claims (8)
1. A key information extraction method, the method comprising:
Acquiring an original document image, and performing data amplification processing on the original document image to obtain an initial training image;
Training a preset lightweight detection model by using the initial training image to obtain a trained lightweight detection model;
Obtaining a certificate image to be identified, and performing target area screening treatment on the certificate image to be identified by using the trained lightweight detection model to obtain a certificate area diagram;
performing angle correction processing on the certificate regional graph to obtain a standard regional graph;
performing target detection processing on the standard region map by using a preset target detection model in any direction to obtain a target rectangular map;
performing text recognition processing on the target rectangular graph to obtain key information;
The target detection processing is performed on the standard area diagram by using a preset target detection model in any direction to obtain a target rectangular diagram, and the target rectangular diagram comprises: carrying out rectangular frame labeling on the relevant areas of the key information in the standard area diagram by using a preset target detection model in any direction; acquiring a twiddle factor corresponding to the marked rectangular frame; judging whether the twiddle factor is a twiddle factor threshold; if the twiddle factor is a twiddle factor threshold, capturing a picture corresponding to the rectangular frame from the standard region picture to obtain a target rectangular picture; if the twiddle factor is not a twiddle factor threshold, carrying out affine transformation on the rectangular frame, and intercepting pictures corresponding to the rectangular frame after affine transformation from the standard area diagram to obtain a target rectangular diagram;
the affine transformation of the rectangular frame comprises: mapping the rectangular frame to a preset two-dimensional rectangular coordinate system, and extracting coordinate values of the rectangular frame; transforming the coordinate values according to a preset affine transformation formula to obtain transformed coordinate values; mapping the transformation coordinate values on the two-dimensional rectangular coordinate system to obtain a rectangular frame after affine transformation;
The preset affine transformation formula is that Wherein/>To change the coordinate value,/>Is coordinate value,/>Is a preset affine change matrix.
2. The method for extracting key information according to claim 1, wherein the performing an angle correction process on the document area map to obtain a standard area map includes:
Performing angle prediction processing on the certificate area diagram by using a preset four-classification model to obtain angle information of the certificate area diagram;
Judging whether the angle information accords with a preset angle standard or not;
If the angle information accords with the preset angle standard, judging that the certificate area diagram is a standard area diagram;
And if the angle information does not accord with the preset angle standard, carrying out angle correction on the certificate area diagram to obtain a standard area diagram.
3. The method for extracting key information according to claim 1, wherein the text recognition processing is performed on the target rectangular chart to obtain the key information, comprising:
performing feature extraction processing on the target rectangular graph to obtain a feature sequence, wherein the feature sequence comprises a plurality of components;
probability calculation is carried out on the components by using a preset activation function, so that probability values of the components are obtained;
and determining the component corresponding to the maximum probability value as key information.
4. The method for extracting key information according to any one of claims 1 to 3, wherein training a preset lightweight detection model by using the initial training image to obtain a trained lightweight detection model comprises:
Performing frame selection processing on the initial training image by using a preset priori frame to obtain a prediction area diagram;
calculating the coincidence value between the predicted area diagram and a preset real area diagram according to a coincidence value formula;
And when the coincidence value is smaller than a preset threshold value, adjusting the internal parameters of the preset lightweight detection model until the coincidence value is larger than or equal to the preset threshold value, and obtaining the trained lightweight detection model.
5. The key information extraction method of claim 4, wherein the coincidence value formula includes:
Wherein, For the coincidence value,/>For the prediction area map,/>For the real region map,/>For the intersection between the prediction area map and the real area map,Is the union between the predicted region map and the real region map.
6. A key information extraction apparatus for implementing the key information extraction method according to any one of claims 1 to 5, characterized in that the apparatus comprises:
the data amplification module is used for acquiring an original document image, and carrying out data amplification processing on the original document image to obtain an initial training image;
The model training module is used for training a preset lightweight detection model by utilizing the initial training image to obtain a trained lightweight detection model;
the regional screening module is used for acquiring a certificate image to be identified, and performing target regional screening treatment on the certificate image to be identified by utilizing the trained lightweight detection model to obtain a certificate regional graph;
The angle correction module is used for performing angle correction processing on the certificate area map to obtain a standard area map;
The target detection module is used for carrying out target detection processing on the standard region graph by utilizing a preset target detection model in any direction to obtain a target rectangular graph;
and the text recognition module is used for carrying out text recognition processing on the target rectangle graph to obtain key information.
7. An electronic device, the electronic device comprising:
At least one processor; and
A memory communicatively coupled to the at least one processor; wherein,
The memory stores instructions executable by the at least one processor to enable the at least one processor to perform the critical information extraction method of any of claims 1 to 5.
8. A computer readable storage medium storing a computer program, wherein the computer program when executed by a processor implements the key information extraction method according to any one of claims 1 to 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011581164.XA CN112668575B (en) | 2020-12-28 | 2020-12-28 | Key information extraction method and device, electronic equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011581164.XA CN112668575B (en) | 2020-12-28 | 2020-12-28 | Key information extraction method and device, electronic equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112668575A CN112668575A (en) | 2021-04-16 |
| CN112668575B true CN112668575B (en) | 2024-05-21 |
Family
ID=75411003
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011581164.XA Active CN112668575B (en) | 2020-12-28 | 2020-12-28 | Key information extraction method and device, electronic equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112668575B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113706422B (en) * | 2021-10-28 | 2022-03-18 | 深圳市亚略特科技股份有限公司 | Image correction method, device, equipment and medium based on key point acquisition |
| CN114596573A (en) * | 2022-03-22 | 2022-06-07 | 中国平安人寿保险股份有限公司 | Birth certificate identification method and device, computer equipment and storage medium |
| CN115375998B (en) * | 2022-10-24 | 2023-03-17 | 成都新希望金融信息有限公司 | Certificate identification method and device, electronic equipment and storage medium |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106909888A (en) * | 2017-01-22 | 2017-06-30 | 南京开为网络科技有限公司 | It is applied to the face key point tracking system and method for mobile device end |
| CN109697440A (en) * | 2018-12-10 | 2019-04-30 | 浙江工业大学 | A kind of ID card information extracting method |
| CN109859101A (en) * | 2019-01-18 | 2019-06-07 | 黑龙江八一农垦大学 | The recognition methods of corps canopy thermal infrared images and system |
| CN110363199A (en) * | 2019-07-16 | 2019-10-22 | 济南浪潮高新科技投资发展有限公司 | Certificate image text recognition method and system based on deep learning |
| CN111027450A (en) * | 2019-12-04 | 2020-04-17 | 深圳市新国都金服技术有限公司 | Bank card information identification method and device, computer equipment and storage medium |
| CN111767859A (en) * | 2020-06-30 | 2020-10-13 | 北京百度网讯科技有限公司 | Image correction method and device, electronic equipment and computer-readable storage medium |
| CN112016547A (en) * | 2020-08-20 | 2020-12-01 | 上海天壤智能科技有限公司 | Image character recognition method, system and medium based on deep learning |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11202079B2 (en) * | 2018-02-05 | 2021-12-14 | Tencent America LLC | Method and apparatus for video decoding of an affine model in an intra block copy mode |
-
2020
- 2020-12-28 CN CN202011581164.XA patent/CN112668575B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106909888A (en) * | 2017-01-22 | 2017-06-30 | 南京开为网络科技有限公司 | It is applied to the face key point tracking system and method for mobile device end |
| CN109697440A (en) * | 2018-12-10 | 2019-04-30 | 浙江工业大学 | A kind of ID card information extracting method |
| CN109859101A (en) * | 2019-01-18 | 2019-06-07 | 黑龙江八一农垦大学 | The recognition methods of corps canopy thermal infrared images and system |
| CN110363199A (en) * | 2019-07-16 | 2019-10-22 | 济南浪潮高新科技投资发展有限公司 | Certificate image text recognition method and system based on deep learning |
| CN111027450A (en) * | 2019-12-04 | 2020-04-17 | 深圳市新国都金服技术有限公司 | Bank card information identification method and device, computer equipment and storage medium |
| CN111767859A (en) * | 2020-06-30 | 2020-10-13 | 北京百度网讯科技有限公司 | Image correction method and device, electronic equipment and computer-readable storage medium |
| CN112016547A (en) * | 2020-08-20 | 2020-12-01 | 上海天壤智能科技有限公司 | Image character recognition method, system and medium based on deep learning |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112668575A (en) | 2021-04-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112699775B (en) | Certificate identification method, device, equipment and storage medium based on deep learning | |
| CN112668575B (en) | Key information extraction method and device, electronic equipment and storage medium | |
| CN112052850B (en) | License plate recognition method and device, electronic equipment and storage medium | |
| CN112861648B (en) | Character recognition method, character recognition device, electronic equipment and storage medium | |
| CN112102402B (en) | Flash light spot position identification method and device, electronic equipment and storage medium | |
| WO2022105179A1 (en) | Biological feature image recognition method and apparatus, and electronic device and readable storage medium | |
| CN112541443B (en) | Invoice information extraction method, invoice information extraction device, computer equipment and storage medium | |
| CN113705462B (en) | Face recognition method, device, electronic equipment and computer readable storage medium | |
| CN110675940A (en) | Pathological image labeling method and device, computer equipment and storage medium | |
| CN113887438B (en) | Watermark detection method, device, equipment and medium for face image | |
| CN112508145A (en) | Electronic seal generation and verification method and device, electronic equipment and storage medium | |
| CN113313114B (en) | Certificate information acquisition method, device, equipment and storage medium | |
| CN112132016B (en) | Bill information extraction method and device and electronic equipment | |
| CN113723280B (en) | Method, device, equipment and medium for detecting countermeasure sample based on static face | |
| CN112862703B (en) | Image correction method and device based on mobile photographing, electronic equipment and medium | |
| CN113627394A (en) | Face extraction method and device, electronic equipment and readable storage medium | |
| CN112560855A (en) | Image information extraction method and device, electronic equipment and storage medium | |
| CN114913518B (en) | License plate recognition method, device, equipment and medium based on image processing | |
| CN115439850B (en) | Method, device, equipment and storage medium for identifying image-text characters based on examination sheets | |
| CN113989548B (en) | Certificate classification model training method and device, electronic equipment and storage medium | |
| CN113177543B (en) | Certificate identification method, device, equipment and storage medium | |
| CN113888086B (en) | Article signing method, device, equipment and storage medium based on image recognition | |
| CN113221888B (en) | License plate number management system test method and device, electronic equipment and storage medium | |
| CN113255456B (en) | Inactive living body detection method, inactive living body detection device, electronic equipment and storage medium | |
| CN115798004A (en) | Face card punching method and device based on local area, electronic equipment and medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |