CN113033530B - Certificate copying detection method and device, electronic equipment and readable storage medium - Google Patents
Certificate copying detection method and device, electronic equipment and readable storage medium Download PDFInfo
- Publication number
- CN113033530B CN113033530B CN202110596941.6A CN202110596941A CN113033530B CN 113033530 B CN113033530 B CN 113033530B CN 202110596941 A CN202110596941 A CN 202110596941A CN 113033530 B CN113033530 B CN 113033530B
- Authority
- CN
- China
- Prior art keywords
- image
- detection
- layer
- detected
- certificate
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/41—Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/22—Image preprocessing by selection of a specific region containing or referencing a pattern; Locating or processing of specific regions to guide the detection or recognition
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/46—Extracting features or characteristics from the video content, e.g. video fingerprints, representative shots or key frames
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computational Linguistics (AREA)
- Multimedia (AREA)
- Software Systems (AREA)
- General Health & Medical Sciences (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Biophysics (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Biomedical Technology (AREA)
- Mathematical Physics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Image Analysis (AREA)
Abstract
The application provides a certificate copying detection method, a certificate copying detection device, electronic equipment and a readable storage medium, and relates to the field of computer image processing. The method comprises the following steps: acquiring video data for certificate copying detection; judging whether an image to be detected exists in an image frame of the video data or not based on a preset detection strategy, wherein the image to be detected is an image needing certificate copying detection; when an image to be detected exists in the image frame, the image to be detected is input into the trained copying detection model based on deep learning, a first detection result of the copying detection model for the image to be detected is obtained, the first detection result represents that the certificate image in the image to be detected is copied or not, and therefore each image frame in video data does not need to be detected, the number of the images detected by the copying detection model is favorably reduced, and the copying detection efficiency of the certificate is improved.
Description
Technical Field
The application relates to the field of computer image processing, in particular to a certificate copying detection method and device, an electronic device and a readable storage medium.
Background
With the popularization of internet finance, services such as opening accounts and handling businesses on line gradually become important businesses in industries such as finance and e-commerce. Current online-based services increase the risk of authenticity of user information compared to offline services. For example, a user may copy credential information on a display screen of a device such as a mobile phone or a computer and upload the credential information as user information. The copied certificate may not be the true certificate of the user himself or may be edited, tampered, forged, etc. Therefore, the business needs to identify the situation of the document to be copied. At present, the copying detection is generally carried out manually for checking and auditing, so that the processing efficiency is low.
Disclosure of Invention
An object of the embodiments of the present application is to provide a method and an apparatus for detecting a document copy, an electronic device, and a readable storage medium, which can improve the efficiency of document copy detection.
In order to achieve the above object, embodiments of the present application are implemented as follows:
in a first aspect, an embodiment of the present application provides a method for detecting a document reproduction, where the method includes:
acquiring video data for certificate copying detection;
judging whether an image to be detected exists in an image frame of the video data or not based on a preset detection strategy, wherein the image to be detected is an image needing certificate copying detection;
when the image to be detected exists in the image frame, the image to be detected is input into a trained deep learning-based copying detection model, a first detection result of the copying detection model on the image to be detected is obtained, and the first detection result represents whether the certificate image in the image to be detected is copied or not.
In the above embodiment, before the duplication detection module performs duplication detection, the image frames of the video data are preprocessed to determine whether an image to be detected exists in the video data, and when the image to be detected exists in the video data, the image to be detected is input into the duplication detection module, and the image to be detected is subjected to duplication detection by the duplication detection module.
With reference to the first aspect, in some optional embodiments, before acquiring the video data for the document duplication detection, the method further comprises:
training the network model through a first type image set and a second type image set to obtain the copying detection model, wherein the first type image set comprises images obtained by copying certificates, and the second type image set comprises images not obtained by copying certificates;
the network model comprises an cbr convolution module, a crc convolution module and a Deep convolution module, wherein the cbr convolution module comprises convolution layers, a batch normalization layer and a Relu activation function layer which are connected in series, the crc convolution module comprises a convolution layer, a Relu activation function layer and a convolution layer, the Deep convolution module comprises at least two cbr convolution modules which are connected in series, two ends of the Relu activation function layer in the crc convolution module are respectively connected with the two convolution layers, and the cbr convolution module, the crc convolution module and the Deep convolution module are used for extracting features of the first type image set and the second type image set.
In the above embodiment, the structure of the network model is simplified, which is beneficial to prompt the computation rate. In addition, the network model is trained to obtain a copying detection model based on deep learning, so that the copying detection model can detect whether the certificate image in the image to be detected is copied, manual detection is replaced by the copying detection model, and the efficiency of certificate copying detection is improved.
With reference to the first aspect, in some optional embodiments, determining whether an image to be detected exists in image frames of the video data based on a preset detection policy includes:
judging whether a first image area representing a display screen and a second image area representing a certificate image exist in each image frame of the video data;
and when the first region and the second region exist in any image frame of the video data and the second region is in the first region, confirming the any image frame as the image to be detected.
In the above embodiments, the copied document image includes an image obtained by copying the document displayed on the display screen. Whether the display screen exists in the image frame and whether the certificate image exists or not are detected, so that whether the image to be detected exists in the image frame of the video data can be quickly determined.
With reference to the first aspect, in some optional embodiments, determining whether a first region representing a display screen and a second region representing a certificate map exist in each image frame of the video data includes:
extracting edge features of any image frame of the video data through an edge detection algorithm;
when a frame of the display screen exists in the edge feature of any image frame, determining that the first region exists in any image frame;
and when the border of the certificate image exists in the edge feature of any image frame, determining that the second region exists in any image frame.
In the above embodiment, the border of the display screen and the certificate can be determined quickly and accurately by the edge detection algorithm, so that whether the image to be detected exists in the image frame of the video data can be determined conveniently and quickly.
With reference to the first aspect, in some optional implementations, when the first region and the second region exist in any image frame of the video data, and the second region is in the first region, determining that the any image frame is the image to be detected includes:
when the first region and the second region exist in any image frame and the second region is in the first region, judging whether the area ratio of the second region to the first region is within a preset range;
and when the area ratio of the first region to the second region is within the preset range, determining that any image frame is the image to be detected.
With reference to the first aspect, in some optional embodiments, the method further comprises:
and when the first detection result shows that the certificate image in the image to be detected is not copied, inputting the image to be detected into a certificate detection model to obtain a second detection result for performing authenticity detection on the image to be detected by the certificate detection model.
With reference to the first aspect, in some optional embodiments, the method further comprises:
and when the first detection result shows that the certificate image in the image to be detected is copied, sending prompt information for representing that the certificate image is copied.
In a second aspect, the present application further provides a model training method, including:
acquiring a first type image set and a second type image set, wherein the first type image set comprises images obtained by copying certificates, and the second type image set comprises images which are not obtained by copying certificates;
through first type of image collection with second type of image collection trains the network model, obtains the reproduction detection model for treating the measured image and carry out certificate reproduction detection, wherein, the network model includes cbr convolution module, crc convolution module and Deep convolution module, cbr convolution module includes the convolution layer, batch standardization layer and the Relu activation function layer that concatenate each other, crc convolution module includes convolution layer, Relu activation function layer and convolution layer, Deep convolution module includes two at least cbr convolution modules that concatenate, among the crc convolution module the both ends of Relu activation function layer respectively with two the convolution layer is connected, cbr convolution module, crc convolution module and Deep convolution module are used for carrying out the feature extraction to first type of image collection with second type of image collection.
In a third aspect, the present application further provides a device for detecting a document reproduction, the device comprising:
the first acquisition unit is used for acquiring video data for certificate copying detection;
the preprocessing unit is used for judging whether an image to be detected exists in the image frames of the video data or not based on a preset detection strategy, and the image to be detected is an image needing certificate copying detection;
and the copying detection unit is used for inputting the image to be detected into a trained copying detection model based on deep learning when the image to be detected exists in the image frame to obtain a first detection result of the copying detection model on the image to be detected, and the first detection result represents that the certificate image in the image to be detected is copied or not.
In a fourth aspect, the present application further provides a model training apparatus, the apparatus comprising:
the second acquisition unit is used for acquiring a first type of image set and a second type of image set, wherein the first type of image set comprises images obtained by copying certificates, and the second type of image set comprises images not obtained by copying the certificates;
the model training unit is used for training a network model through the first type of image set and the second type of image set to obtain a copying detection model and used for carrying out certificate copying detection on an image to be detected, wherein the network model comprises an cbr convolution module, a crc convolution module and a Deep convolution module, the cbr convolution module comprises convolution layers, a batch normalization layer and a Relu activation function layer which are connected in series, the crc convolution module comprises a convolution layer, a Relu activation function layer and a convolution layer, the Deep convolution module comprises at least two cbr convolution modules which are connected in series, two ends of the Relu activation function layer in the crc convolution module are respectively connected with the two convolution layers, and the cbr convolution module, the crc convolution module and the Deep convolution module are used for carrying out feature extraction on the first type of image set and the second type of image set.
In a fifth aspect, the present application further provides an electronic device, which includes a processor and a memory coupled to each other, wherein the memory stores a computer program, and when the computer program is executed by the processor, the electronic device is caused to perform the method described above.
In a sixth aspect, the present application also provides a computer-readable storage medium having stored thereon a computer program which, when run on a computer, causes the computer to perform the above-mentioned method.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings that are required to be used in the embodiments of the present application will be briefly described below, it should be understood that the following drawings only illustrate some embodiments of the present application and therefore should not be considered as limiting the scope, and that those skilled in the art can also obtain other related drawings based on the drawings without inventive efforts.
Fig. 1 is a schematic flow chart of a certificate copying detection method provided in an embodiment of the present application.
Fig. 2 is a schematic network structure diagram of a copy detection model provided in the embodiment of the present application.
Fig. 3a is a schematic diagram of a network structure of an cbr convolution module according to an embodiment of the present application.
Fig. 3b is a schematic diagram of a network structure of a crc convolution module according to an embodiment of the present application.
Fig. 3c is a schematic network structure diagram of the Deep convolution module according to the embodiment of the present application.
Fig. 4 is a block diagram of a certificate duplication detection apparatus according to an embodiment of the present application.
Fig. 5 is a schematic flowchart of a model training method according to an embodiment of the present application.
Fig. 6 is a block diagram of a model training apparatus according to an embodiment of the present application.
Icon: 200-certificate reproduction detection means; 210-a first obtaining unit; 220-a pre-processing unit; 230-a reproduction detection unit; 400-a model training device; 410-a second obtaining unit; 420-model training unit.
Detailed Description
The technical solutions in the embodiments of the present application will be described below with reference to the drawings in the embodiments of the present application. It should be noted that the terms "first," "second," and the like are used merely to distinguish one description from another, and are not intended to indicate or imply relative importance.
First embodiment
The application provides an electronic equipment can be used for carrying out certificate reproduction and detection. The electronic device may include a processing module and a memory module. The storage module stores a computer program which, when executed by the processing module, enables the electronic device to perform the steps of the methods described below.
The electronic device may further include other modules, for example, the electronic device may further include a communication module for performing data interaction with other devices. For example, when the electronic device is a user terminal, the user terminal may establish a communication connection with the server through the communication module to perform data interaction.
The processing module, the storage module and the communication module are electrically connected directly or indirectly to realize data transmission or interaction. For example, the components may be electrically connected to each other via one or more communication buses or signal lines.
Referring to fig. 1, the present application further provides a certificate duplication detection method, which can be applied to the electronic device, and is executed or implemented by the electronic device. Wherein, the method can comprise the following steps:
step S110, acquiring video data for certificate copying detection;
step S120, judging whether an image to be detected exists in the image frames of the video data based on a preset detection strategy, wherein the image to be detected is an image needing certificate copying detection;
step S130, when the image to be detected exists in the image frame, inputting the image to be detected into a trained deep learning-based reproduction detection model to obtain a first detection result of the reproduction detection model on the image to be detected, wherein the first detection result represents whether a certificate image in the image to be detected is reproduced or not.
In this embodiment, before the duplication detection model carries out the duplication detection, carry out the preliminary treatment to video data's image frame earlier to judge whether there is the image that awaits measuring in the video data, and when there is the image that awaits measuring in the video data, just with the image input duplication detection model that awaits measuring, treat the image that awaits measuring by the duplication detection model and carry out the duplication detection, so, need not to detect every image frame in the video data, be favorable to reducing the image quantity that the duplication detection model detected, thereby improve certificate duplication detection's efficiency.
The individual steps of the process are explained in detail below, as follows:
in step S110, the electronic device may capture video data in real time through a camera, or acquire a pre-stored recorded video file from its own storage module or other devices as the video data acquired by the electronic device. The document to be detected may be, but is not limited to, an identification card, a driving license, etc. The camera may be a device of the electronic device itself, or an external module, and is not limited in this respect.
For example, the electronic device is a user terminal, and when a user needs to upload an identity card to the server through the user terminal to perform identity authentication, the user can open a camera on the user terminal and shoot the identity card through the camera to obtain video data. Compared with the method of only uploading a single image for copying detection, the method has the advantages that the video data usually comprises richer environment information and behavior information (such as the user moving a certificate), and the detection accuracy of the screen and the identity card is improved.
It should be noted that the video data may be obtained by the camera copying a certificate image displayed on a display screen of another device, or obtained by the camera directly shooting a certificate, where a source manner of the video data is not particularly limited.
In step S120, a preset detection policy is used to determine whether an image to be detected exists in an image frame of the video data. The image to be detected is the image needing to be subjected to copying detection. The image needing to be subjected to copying detection is the image possibly obtained by copying the certificate. For example, the image frame has a display screen and a certificate region, and the certificate region is included in the display screen region, so that the image frame can be used as an image to be measured.
In step S130, the duplication detection model is obtained by training a deep learning-based network model, and can be used to detect whether the image to be detected is a duplication. And the copying detection model is used for replacing manual detection, so that the detection efficiency can be improved.
When the copying detection model detects the image to be detected and the obtained detection result shows that the certificate image is copied, the video data of the certificate shot by the user has the problem of incredibility, and the electronic equipment can prompt the user to shoot the certificate again until the certificate image in the video data is detected not to be copied.
When the obtained detection result shows that the certificate image is not copied, the video data of the certificate shot by the user is credible video data, and then the subsequent certificate authenticity detection can be carried out.
In this embodiment, before the copy detection model performs the document copy detection, model training is required. That is, before step S110, the method may further include:
training the network model through a first type image set and a second type image set to obtain the copying detection model, wherein the first type image set comprises images obtained by copying certificates, and the second type image set comprises images not obtained by copying certificates. The model training process can be seen in the second embodiment described below.
Referring to fig. 2, fig. 3a, fig. 3b, and fig. 3c, the network model may include cbr convolution modules, a crc convolution module, and a Deep convolution module, wherein the cbr convolution module includes a convolution layer conv, a batch normalization layer bn, and a Relu activation function layer connected in series, the crc convolution module includes a convolution layer conv, a Relu activation function layer, and a convolution layer conv, and two ends of the Relu activation function layer in the crc convolution module are respectively connected to the two convolution layers conv. The Deep convolution module comprises at least two cbr convolution modules which are connected in series, wherein the cbr convolution module, the crc convolution module and the Deep convolution module are used for extracting features of the first type image set and the second type image set.
In the network model, the convolutional layer may be denoted by "conv", and the batch normalization layer may be denoted by "bn", and the reduced Linear Unit (modified Linear Unit) activation function may be denoted by "Relu". "cbr" is the initial composition of "conv", "bn" and "relu"; "crc" is the initial composition of "conv", "relu" and "conv".
In this embodiment, the definition of the parameters in the mathematical formulas corresponding to the cbr convolution module, the crc convolution module, and the Deep convolution module can be as follows:
the cbr convolution module, the crc convolution module, and the Deep convolution module may have the following formulas:
other convolutional layer modules may be included in the network model, for example, the network model may further include:
、
、…、
、
、
、
、
、
、
. The convolutional layer modules are mutually matched and used for enabling the network model to perform feature extraction of image data.
In fig. 2, the convolution module of the network structure may include an operation unit 5, an operation unit 6, …, an operation unit 26, and an operation unit 27. The operation formula of the operation unit n may be the following formula (n), and n is any integer of 5 to 27. For example, the operation formula of the operation unit 5 is the following formula (5). In fig. 4, input data of the arithmetic unit indicated by an arrow is an output parameter of the arithmetic unit at the end of the arrow. For example, the parameter y1 output by the arithmetic unit 5 is an input parameter of the arithmetic unit 6. The mathematical formula corresponding to the corresponding operation unit of the convolutional layer module can be as follows:
in the above formula, the meaning of each type of parameter is as follows:
x represents an image input to the duplication detection model;
y represents data output by the duplication detection model, such as duplication detection results of certificate images in video data;
the subscript mean means calculating an arithmetic mean;
max refers to the calculated maximum;
the subscript a has no other meaning and is used for distinguishing from the subscripts b and c;
the subscript b is the abbreviation of English "batch", which refers to each batch of data in the batch processing;
the subscript c is abbreviated as "channel" and refers to a channel of the image, for example, taking a color image as an example, the channel of the image includes a Red (R, R) channel, a Green (G) channel, and a Blue (B) channel;
the subscript w is the abbreviation of English width, which means the width of the input image;
the subscript h is short for English "height", which means the height of the input image;
the subscript cat indicates the tensor splicing operation, corresponding to the mathematical symbol "
”。
In the above formula, the values of b, c, w and h are all integers, and the subscript i is the corresponding integer in the formula, which is well known to those skilled in the art. For example, in the above formula (18), i may be any integer of 0 to h 4. b. The subscripts of c, w and h are the digital subscripts of 'y' in the corresponding formula and are used for distinguishing data of different convolutional layers.
In the above embodiment, the network model is used for training to obtain the copying detection model based on deep learning, so that the copying detection model can detect whether the certificate image in the image to be detected is copied, and the copying detection model replaces manual detection, thereby being beneficial to improving the efficiency of certificate copying detection.
In this embodiment, the preset detection strategy may be determined according to actual conditions. As an alternative implementation, step S120 may include:
judging whether a first image area representing a display screen and a second image area representing a certificate image exist in each image frame of the video data;
and when the first region and the second region exist in any image frame of the video data and the second region is in the first region, confirming the any image frame as the image to be detected.
The method for judging whether a first region representing a display screen and a second region representing a certificate image exist in each image frame of the video data comprises the following steps:
extracting edge features of any image frame of the video data through an edge detection algorithm;
when a frame of the display screen exists in the edge feature of any image frame, determining that the first region exists in any image frame;
and when the border of the certificate image exists in the edge feature of any image frame, determining that the second region exists in any image frame.
In other embodiments, the preset detection policy may be a YOLO (a target detection algorithm) algorithm, which is a target detection model based on deep learning, and may be used to find some specific objects in a picture, where the target detection model may identify the types of the objects and mark the frame positions of the objects in the image, for example, the YOLO algorithm may be applied to face recognition, and may locate the face positions.
The electronic equipment can detect whether a display screen region of other equipment and a certificate map exist in an image frame of video data through a YOLO algorithm, and when the display screen region and the certificate map are determined to exist, a first region of the display screen region and a second region of the certificate map are positioned so as to determine whether the image frame is an image to be detected.
Understandably, the image frames obtained by copying the images of the documents displayed on the display screens of other devices usually include the regions of the copied display screens and the regions of the documents displayed on the display screens. Based on the image characteristics of the display screen and the image area of the certificate, whether the display screen exists in the image frame and whether the certificate image exists in the image frame can be determined through an edge detection algorithm or a YOLO algorithm, and therefore whether the image to be detected exists in the image frame of the video data can be determined rapidly.
And if the fact that the display screen does not exist in the image frame and the certificate image does not exist is determined through an edge detection algorithm or a YOLO algorithm, directly detecting. And if the image frame is determined to have no display screen but a certificate image, determining that the image frame is not the certificate reproduction image, and at the moment, inputting the image frame into a reproduction detection model for reproduction detection.
Wherein, when the first region and the second region exist in any image frame of the video data and the second region is in the first region, determining the any image frame as the image to be detected comprises:
when the first region and the second region exist in any image frame and the second region is in the first region, judging whether the area ratio of the second region to the first region is within a preset range;
and when the area ratio of the first region to the second region is within the preset range, determining that any image frame is the image to be detected.
The preset range can be set according to actual conditions, and can be 30% -80%, for example. The method comprises the steps of calculating the area ratio of a certificate image area (second image area) to a display screen image area (first image area), judging whether the obtained area ratio is within a preset range, and if the obtained area ratio is within the preset range, determining that the image frame is an image to be detected and indicating that the image frame is possibly a certificate reproduction image.
If the area ratio is larger than the maximum value of the preset range, the detection result indicating that the image frame is the certificate reproduction image can be directly output, and the image frame does not need to be input into a reproduction detection model for reproduction detection.
If the area ratio is smaller than the minimum value of the preset range, the detection result indicating that the image frame is not the certificate reproduction image can be directly output, and the image frame does not need to be input into a reproduction detection model for reproduction detection.
In the embodiment, the real-time video data which is shot after the user opens the camera can be used as input data, and the images which are obtained by the camera are transmitted in a time sequence mode, so that a series of behavior information which is ready to be shot by the user and the surrounding environment information of richer certificate images can be obtained, and the problem that the images uploaded by the user lack environment information due to the fact that the images are too close to the certificates, and therefore the judgment of copying is not facilitated is solved.
In addition, the coincidence degree of the certificate and the screen is combined with real-time video data, and the change of the coincidence degree at different time points is compared to provide a basis for judging the reproduction. The real-time video data contains information about the behavior of the user-adjusted document, so that a change in the degree of overlap indicates that the document is not on the screen. Meanwhile, the change of the coincidence degree can be comprehensively judged by combining the size of the coincidence area. If the coincidence degree changes, but the coincidence area is large, the situation that the user puts the certificate above the screen can be possibly, the user still can judge that the certificate is not copied, and therefore the accuracy of copying judgment can be improved.
As an optional implementation, the method may further include:
and when the first detection result shows that the certificate image in the image to be detected is not copied, inputting the image to be detected into a certificate detection model to obtain a second detection result for performing authenticity detection on the image to be detected by the certificate detection model.
The certificate detection model detection can be used for performing authenticity detection on user information in a certificate diagram, and the authenticity detection mode is well known to those skilled in the art and is not described herein again. In this embodiment, through combination of reproduction detection and authenticity detection, the accuracy and reliability of the uploaded certificate image detection can be improved.
In this embodiment, when it is determined that the certificate image in the video data is not a copy, at this time, an image frame with the certificate image and the highest definition and without a block in the certificate image area may be selected from the video data to be used as a target image, and the target image is input into the certificate detection model, so that the accuracy of authenticity detection is improved. The definition determination method and the occlusion determination method are well known to those skilled in the art, and are not described herein again.
Similarly, when the image to be detected exists in the video data, the electronic equipment can select the image frame which has the certificate image and the highest definition and is not shielded in the certificate image area as the target image, and the target image is input into the copying detection model, so that the accuracy and the reliability of copying detection are improved.
It should be noted that the method may be executed by a single side of the user terminal or the server, or by the user terminal interacting with the server. When the image to be detected is interactively executed, the user terminal may execute steps S110 and S120, and then the user terminal sends the determined image to be detected to the server, and when the server receives the image to be detected, the server inputs the image to be detected to the copying detection model, and then starts to execute step S130, so that the amount of computation of the copying detection of the server may be reduced, and the computation pressure of the server may be reduced.
As an optional implementation, the method may further include:
and when the first detection result shows that the certificate image in the image to be detected is copied, sending prompt information for representing that the certificate image is copied.
In this embodiment, when it is determined that the image to be detected is a copied image, the electronic device sends prompt information to prompt the user that the currently uploaded certificate image is the copied image, so that the user is required to adjust the uploaded video data in time, and the user is prevented from uploading the certificate image by copying.
Referring to fig. 4, the present application further provides a device 200 for detecting a document reproduction, which can be applied to the electronic device described above for executing the steps of the method. The document copying detection device 200 includes at least one software function module which can be stored in a memory module in the form of software or Firmware (Firmware) or solidified in an Operating System (OS) of the electronic device. The processing module is used for executing executable modules stored in the storage module, such as software functional modules and computer programs included in the certificate duplication detection apparatus 200.
The device 200 can include a first capture unit 210, a preprocessing unit 220, and a duplication detection unit 230, and can perform the following steps:
a first acquisition unit 210 for acquiring video data for certificate duplication detection;
the preprocessing unit 220 is configured to determine whether an image to be detected exists in image frames of the video data based on a preset detection strategy, where the image to be detected is an image that needs to be subjected to certificate copying detection;
and the copying detection unit 230 is configured to, when the image frame has the image to be detected, input the image to be detected into a trained deep learning-based copying detection model to obtain a first detection result of the copying detection model on the image to be detected, where the first detection result represents whether the certificate image in the image to be detected is copied or not.
Optionally, the credential duplication detection device 200 may further include a model training unit configured to, before the first acquisition unit 210 performs step S110:
training the network model through a first type image set and a second type image set to obtain the copying detection model, wherein the first type image set comprises images obtained by copying certificates, and the second type image set comprises images not obtained by copying certificates;
the network model comprises an cbr convolution module, a crc convolution module and a Deep convolution module, wherein the cbr convolution module comprises convolution layers, a batch normalization layer and a Relu activation function layer which are connected in series, the crc convolution module comprises a convolution layer, a Relu activation function layer and a convolution layer, the Deep convolution module comprises at least two cbr convolution modules which are connected in series, two ends of the Relu activation function layer in the crc convolution module are respectively connected with the two convolution layers, and the cbr convolution module, the crc convolution module and the Deep convolution module are used for extracting features of the first type image set and the second type image set.
Optionally, the preprocessing unit 220 may be configured to: judging whether a first image area representing a display screen and a second image area representing a certificate image exist in each image frame of the video data; and when the first region and the second region exist in any image frame of the video data and the second region is in the first region, confirming the any image frame as the image to be detected.
Optionally, the preprocessing unit 220 may be further configured to:
extracting edge features of any image frame of the video data through an edge detection algorithm;
when a frame of the display screen exists in the edge feature of any image frame, determining that the first region exists in any image frame;
and when the border of the certificate image exists in the edge feature of any image frame, determining that the second region exists in any image frame.
Optionally, the preprocessing unit 220 may be further configured to:
when the first region and the second region exist in any image frame and the second region is in the first region, judging whether the area ratio of the second region to the first region is within a preset range;
and when the area ratio of the first region to the second region is within the preset range, determining that any image frame is the image to be detected.
Optionally, the certificate copying detection apparatus 200 may further include a authenticity detection unit, configured to, when the first detection result indicates that the certificate diagram in the image to be detected is not copied, input the image to be detected into a certificate detection model, and obtain a second detection result that the certificate detection model performs authenticity detection on the image to be detected.
Optionally, the certificate copying detection apparatus 200 may further include a prompting unit, configured to send a prompting message indicating that the certificate image is copied when the first detection result indicates that the certificate image in the image to be detected is copied.
It should be clearly understood by those skilled in the art that, for convenience and brevity of description, the specific working processes of the electronic device and the certificate reproduction detecting apparatus 200 described above may refer to the corresponding processes of the steps in the foregoing method, and will not be described in detail herein.
Second embodiment
Referring to fig. 5, the present application further provides a model training method, which can be applied to the electronic device described above, and is executed by the electronic device or used for implementing the steps of the method. The method may comprise the steps of:
step S310, a first type image set and a second type image set are obtained, wherein the first type image set comprises images obtained by copying certificates, and the second type image set comprises images not obtained by copying certificates;
step S310, training a network model through the first type image set and the second type image set to obtain a copying detection model for carrying out certificate copying detection on an image to be detected, wherein the network model comprises an cbr convolution module, a crc convolution module and a Deep convolution module, the cbr convolution module comprises convolution layers, a batch normalization layer and a Relu activation function layer which are connected in series, the crc convolution module comprises convolution layers, a Relu activation function layer and convolution layers, the Deep convolution module comprises at least two cbr convolution modules which are connected in series, two ends of the Relu activation function layer in the crc convolution module are respectively connected with the two convolution layers, and the cbr convolution module, the crc convolution module and the Deep convolution module are used for carrying out feature extraction on the first type image set and the second type image set.
In step S310, the user may prepare a large number of images obtained by copying the document and images not obtained by copying the document in advance as the first type image set and the second type image set respectively. When the model training is performed, the prepared first-type image set and second-type image set can be directly obtained from corresponding storage addresses or storage devices.
Understandably, the second type of image set may be a set of images obtained by copying the certificate image displayed in the display screen of the other device, the second type of image set may be a set of images obtained by directly shooting the certificate, and both the first type of image set and the second type of image set serve as training images of the training model. The number of images in the first-type image set and the second-type image set can be determined according to practical situations, and is not particularly limited herein.
The electronic device may train a network model as shown in fig. 2 by using the first type of image set and the second type of image set, thereby obtaining a trained network model. And then, the electronic equipment tests and verifies the trained network model by using the verification image set, so that the network model subjected to test and verification can be obtained. The accuracy and reliability of the copying detection of the image to be detected can be improved through the network model subjected to test verification, and the network model subjected to test verification is the copying detection model.
The network model forms a copying detection model capable of effectively identifying copying of the certificate by learning data of the certificate image under various environments, including but not limited to training images under different light rays, training images with different background textures, training images obtained by placing the certificate on a display screen, training images obtained by copying certificate images displayed by different types of display screens and the like. For example, when the user places the identification card on the screen, the user shoots the identification card under the condition to obtain video data, and when the user conducts copying detection, the copying detection model of the electronic device can still judge that the identification card is not copied.
In this embodiment, the structure of the copying detection model is simplified, so that the structure of the network model is simpler, the calculation speed is improved, the network model is conveniently deployed in a server or various edge devices (such as a personal computer), and the rapid calculation of the certificate copying detection is realized.
Referring to fig. 6, the present application further provides a model training apparatus 400, which can be applied to the electronic device described above for executing the steps of the model training method. The model training apparatus 400 includes at least one software function module which can be stored in a memory module in the form of software or Firmware (Firmware) or solidified in an Operating System (OS) of an electronic device.
The model training apparatus 400 may include a second obtaining unit 410 and a model training unit 420, and may perform the following operation steps:
a second obtaining unit 410, configured to obtain a first type image set and a second type image set, where the first type image set includes images obtained by copying documents, and the second type image set includes images not obtained by copying documents;
the model training unit 420 is used for training a network model through the first type of image set and the second type of image set to obtain a copying detection model and used for carrying out certificate copying detection on an image to be detected, wherein the network model comprises an cbr convolution module, a crc convolution module and a Deep convolution module, the cbr convolution module comprises convolution layers, a batch normalization layer and a Relu activation function layer which are connected in series, the crc convolution module comprises a convolution layer, a Relu activation function layer and a convolution layer, the Deep convolution module comprises at least two cbr convolution modules which are connected in series, two ends of the Relu activation function layer in the crc convolution module are respectively connected with the two convolution layers, and the cbr convolution module, the crc convolution module and the Deep convolution module are used for carrying out feature extraction on the first type of image set and the second type of image set.
In this embodiment, the processing module may be an integrated circuit chip having signal processing capability. The processing module may be a general purpose processor. For example, the Processor may be a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), a Network Processor (NP), or the like; the method, the steps and the logic block diagram disclosed in the embodiments of the present Application may also be implemented or executed by a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other Programmable logic device, a discrete Gate or transistor logic device, or a discrete hardware component.
The memory module may be, but is not limited to, a random access memory, a read only memory, a programmable read only memory, an erasable programmable read only memory, an electrically erasable programmable read only memory, and the like. In this embodiment, the storage module may be configured to store video data, an image to be detected, a reproduction detection model, a detection result, and the like. Of course, the storage module may also be used to store other programs, and the processing module executes the programs after receiving the execution instruction.
The communication module is used for establishing communication connection between the electronic equipment and other equipment through a network and receiving and transmitting data through the network.
It should be noted that, as will be clear to those skilled in the art, for convenience and brevity of description, the specific working process of the model training apparatus 400 described above may refer to the corresponding process of each step in the foregoing method, and will not be described in detail herein.
The embodiment of the application also provides a computer readable storage medium. The computer-readable storage medium has stored therein a computer program which, when run on a computer, causes the computer to perform the method for document reproduction detection as described in the above embodiments, or the method for model training as described above.
From the above description of the embodiments, it is clear to those skilled in the art that the present application can be implemented by hardware, or by software plus a necessary general hardware platform, and based on such understanding, the technical solution of the present application can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a usb disk, a removable hard disk, etc.), and includes several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the method described in the embodiments of the present application.
In summary, the present application provides a certificate copying detection method, device, electronic device and readable storage medium, the method comprising: acquiring video data for certificate copying detection; judging whether an image to be detected exists in an image frame of the video data or not based on a preset detection strategy, wherein the image to be detected is an image needing certificate copying detection; when an image to be detected exists in the image frame, the image to be detected is input into the trained copying detection model based on deep learning, a first detection result of the copying detection model for the image to be detected is obtained, and the first detection result represents that the certificate image in the image to be detected is copied or not. In this scheme, before the detection model that reprints carries out the reproduction detection, carry out the preliminary treatment to video data's image frame earlier to judge whether have the image that awaits measuring in the video data, and when having the image that awaits measuring in the video data, just with the image input reproduction detection model that awaits measuring, treat the image that awaits measuring by the reproduction detection model and carry out the reproduction detection, so, need not to detect every image frame in the video data, be favorable to reducing the image quantity that the reproduction detection model detected, thereby improve the efficiency that the certificate was reprinted and is detected.
In the embodiments provided in the present application, it should be understood that the disclosed apparatus, system, and method may be implemented in other ways. The apparatus, system, and method embodiments described above are illustrative only, as the flowcharts and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present application. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions. In addition, functional modules in the embodiments of the present application may be integrated together to form an independent part, or each module may exist separately, or two or more modules may be integrated to form an independent part.
The above description is only an example of the present application and is not intended to limit the scope of the present application, and various modifications and changes may be made by those skilled in the art. Any modification, equivalent replacement, improvement and the like made within the spirit and principle of the present application shall be included in the protection scope of the present application.
Claims (8)
1. A method of document reproduction detection, the method comprising:
acquiring video data for certificate copying detection;
judging whether an image to be detected exists in an image frame of the video data or not based on a preset detection strategy, wherein the image to be detected is an image needing certificate copying detection;
when the image to be detected exists in the image frame, inputting the image to be detected into a trained deep learning-based copying detection model to obtain a first detection result of the copying detection model on the image to be detected, wherein the first detection result represents whether a certificate image in the image to be detected is copied or not;
the method for judging whether an image to be detected exists in an image frame of the video data based on a preset detection strategy comprises the following steps:
judging whether a first image area representing a display screen and a second image area representing a certificate image exist in each image frame of the video data;
when the first region and the second region exist in any image frame of the video data and the second region is in the first region, confirming the any image frame as the image to be detected;
the network model of the reproduction detection model comprises a layer1、layer2、layer3、layer4、layer5、layer6、layer7、layer8、layer9、layer10、layer11、layer12、layer13、
、
、
、
、
、
The method is used for extracting the characteristics of the image frame, and the corresponding formula is as follows:
wherein, prior to acquiring video data for credential duplication detection, the method further comprises:
training a network model through a first type image set and a second type image set to obtain the copying detection model, wherein the first type image set comprises images obtained by copying certificates, and the second type image set comprises images not obtained by copying certificates;
the network model comprises cbr convolution modules, a crc convolution module and a Deep convolution module, wherein the cbr convolution module comprises convolution layers, a batch normalization layer and a Relu activation function layer which are connected in series, the crc convolution module comprises a convolution layer, a Relu activation function layer and a convolution layer, the Deep convolution module comprises at least two cbr convolution modules which are connected in series, two ends of the Relu activation function layer in the crc convolution module are respectively connected with the two convolution layers, and the cbr convolution module, the crc convolution module and the Deep convolution module are used for extracting features of the first type image set and the second type image set;
in the formula, x represents an image input to the duplication detection model;
y represents data output by the copying detection model;
the subscript mean means calculating an arithmetic mean;
max refers to the calculated maximum;
the footmark a is used for distinguishing the footmarks b and c;
the subscript b indicates each batch of data in the batch processing;
the subscript c refers to the channel of the image;
the subscript w indicates the width of the input image;
the subscript h indicates the height of the input image;
the subscript cat indicates tensor splicing operation;
convcbr() The cbr convolution module is used for carrying out convolution operation on the elements in the brackets;
convdeep() The method is characterized in that the Deep convolution module is used for carrying out convolution operation on elements in brackets;
convcrc() The convolution operation is performed on the elements in the brackets by the crc convolution module.
2. The method of claim 1, wherein determining whether a first region characterizing a display screen and a second region characterizing a credential graph are present in each image frame of the video data comprises:
extracting edge features of any image frame of the video data through an edge detection algorithm;
when a frame of the display screen exists in the edge feature of any image frame, determining that the first region exists in any image frame;
and when the border of the certificate image exists in the edge feature of any image frame, determining that the second region exists in any image frame.
3. The method of claim 1, wherein when the first region and the second region are present in any image frame of the video data and the second region is in the first region, confirming the any image frame as the image to be tested comprises:
when the first region and the second region exist in any image frame and the second region is in the first region, judging whether the area ratio of the second region to the first region is within a preset range;
and when the area ratio of the first region to the second region is within the preset range, determining that any image frame is the image to be detected.
4. The method of claim 1, further comprising:
and when the first detection result shows that the certificate image in the image to be detected is not copied, inputting the image to be detected into a certificate detection model to obtain a second detection result for performing authenticity detection on the image to be detected by the certificate detection model.
5. The method of claim 1, further comprising:
and when the first detection result shows that the certificate image in the image to be detected is copied, sending prompt information for representing that the certificate image is copied.
6. A document reproduction detection device, the device comprising:
the first acquisition unit is used for acquiring video data for certificate copying detection;
the preprocessing unit is used for judging whether an image to be detected exists in the image frames of the video data or not based on a preset detection strategy, and the image to be detected is an image needing certificate copying detection;
the copying detection unit is used for inputting the image to be detected into a trained copying detection model based on deep learning when the image to be detected exists in the image frame to obtain a first detection result of the copying detection model on the image to be detected, wherein the first detection result represents whether a certificate image in the image to be detected is copied or not;
wherein the preprocessing unit is further configured to: judging whether a first image area representing a display screen and a second image area representing a certificate image exist in each image frame of the video data; when the first region and the second region exist in any image frame of the video data and the second region is in the first region, confirming the any image frame as the image to be detected;
the network model of the reproduction detection model comprises a layer1、layer2、layer3、layer4、layer5、layer6、layer7、layer8、layer9、layer10、layer11、layer12、layer13、
、
、
、
、
、
The method is used for extracting the characteristics of the image frame, and the corresponding formula is as follows:
wherein the apparatus further comprises a model training unit, prior to acquiring video data for credential duplication detection, the model training unit further to:
training a network model through a first type image set and a second type image set to obtain the copying detection model, wherein the first type image set comprises images obtained by copying certificates, and the second type image set comprises images not obtained by copying certificates;
the network model comprises cbr convolution modules, a crc convolution module and a Deep convolution module, wherein the cbr convolution module comprises convolution layers, a batch normalization layer and a Relu activation function layer which are connected in series, the crc convolution module comprises a convolution layer, a Relu activation function layer and a convolution layer, the Deep convolution module comprises at least two cbr convolution modules which are connected in series, two ends of the Relu activation function layer in the crc convolution module are respectively connected with the two convolution layers, and the cbr convolution module, the crc convolution module and the Deep convolution module are used for extracting features of the first type image set and the second type image set;
in the formula, x represents an image input to the duplication detection model;
y represents data output by the copying detection model;
the subscript mean means calculating an arithmetic mean;
max refers to the calculated maximum;
the footmark a is used for distinguishing the footmarks b and c;
the subscript b indicates each batch of data in the batch processing;
the subscript c refers to the channel of the image;
the subscript w indicates the width of the input image;
the subscript h indicates the height of the input image;
the subscript cat indicates tensor splicing operation;
convcbr() The cbr convolution module is used for carrying out convolution operation on the elements in the brackets;
convdeep() The method is characterized in that the Deep convolution module is used for carrying out convolution operation on elements in brackets;
convcrc() The convolution operation is performed on the elements in the brackets by the crc convolution module.
7. An electronic device, characterized in that the electronic device comprises a processor and a memory coupled to each other, the memory storing a computer program which, when executed by the processor, causes the electronic device to perform the method according to any of claims 1-5.
8. A computer-readable storage medium, in which a computer program is stored which, when run on a computer, causes the computer to carry out the method according to any one of claims 1-5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110596941.6A CN113033530B (en) | 2021-05-31 | 2021-05-31 | Certificate copying detection method and device, electronic equipment and readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110596941.6A CN113033530B (en) | 2021-05-31 | 2021-05-31 | Certificate copying detection method and device, electronic equipment and readable storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113033530A CN113033530A (en) | 2021-06-25 |
| CN113033530B true CN113033530B (en) | 2022-02-22 |
Family
ID=76455905
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110596941.6A Active CN113033530B (en) | 2021-05-31 | 2021-05-31 | Certificate copying detection method and device, electronic equipment and readable storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113033530B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116168038B (en) * | 2023-04-26 | 2023-06-30 | 创新奇智(青岛)科技有限公司 | Image reproduction detection method and device, electronic equipment and storage medium |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104915649A (en) * | 2015-06-04 | 2015-09-16 | 南京理工大学 | Living person detection method applied to face recognition |
| EP3522053A1 (en) * | 2018-02-06 | 2019-08-07 | Idemia Identity & Security France | Method for authenticating a face |
| CN111079687A (en) * | 2019-12-26 | 2020-04-28 | 京东数字科技控股有限公司 | Certificate camouflage identification method, device, equipment and storage medium |
| CN111291586A (en) * | 2018-12-06 | 2020-06-16 | 北京市商汤科技开发有限公司 | Living body detection method, living body detection device, electronic apparatus, and computer-readable storage medium |
| CN111695406A (en) * | 2020-04-23 | 2020-09-22 | 西安电子科技大学 | Face recognition anti-spoofing method, system and terminal based on infrared ray |
| CN111754441A (en) * | 2020-06-29 | 2020-10-09 | 国网甘肃省电力公司电力科学研究院 | Passive detection method for image copy-paste forgery |
| CN112257685A (en) * | 2020-12-08 | 2021-01-22 | 成都新希望金融信息有限公司 | Face copying recognition method and device, electronic equipment and storage medium |
| CN112528998A (en) * | 2021-02-18 | 2021-03-19 | 成都新希望金融信息有限公司 | Certificate image processing method and device, electronic equipment and readable storage medium |
| CN112580621A (en) * | 2020-12-24 | 2021-03-30 | 成都新希望金融信息有限公司 | Identity card copying and identifying method and device, electronic equipment and storage medium |
Family Cites Families (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8144368B2 (en) * | 1998-01-20 | 2012-03-27 | Digimarc Coporation | Automated methods for distinguishing copies from original printed objects |
| CN108573202A (en) * | 2017-03-17 | 2018-09-25 | 北京旷视科技有限公司 | Identity identifying method, device and system and terminal, server and storage medium |
| CN106991451A (en) * | 2017-04-14 | 2017-07-28 | 武汉神目信息技术有限公司 | A kind of identifying system and method for certificate picture |
| CN111881707B (en) * | 2019-12-04 | 2021-09-14 | 马上消费金融股份有限公司 | Image reproduction detection method, identity verification method, model training method and device |
| CN111275685B (en) * | 2020-01-20 | 2024-06-11 | 中国平安人寿保险股份有限公司 | Method, device, equipment and medium for identifying flip image of identity document |
| CN111767828B (en) * | 2020-06-28 | 2023-12-08 | 京东科技控股股份有限公司 | Certificate image reproduction identification method and device, electronic equipment and storage medium |
| CN112085094B (en) * | 2020-09-08 | 2024-04-05 | 中国平安财产保险股份有限公司 | Document image reproduction detection method, device, computer equipment and storage medium |
| CN112507923B (en) * | 2020-12-16 | 2023-10-31 | 平安银行股份有限公司 | Certificate copying detection method and device, electronic equipment and medium |
-
2021
- 2021-05-31 CN CN202110596941.6A patent/CN113033530B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104915649A (en) * | 2015-06-04 | 2015-09-16 | 南京理工大学 | Living person detection method applied to face recognition |
| EP3522053A1 (en) * | 2018-02-06 | 2019-08-07 | Idemia Identity & Security France | Method for authenticating a face |
| CN111291586A (en) * | 2018-12-06 | 2020-06-16 | 北京市商汤科技开发有限公司 | Living body detection method, living body detection device, electronic apparatus, and computer-readable storage medium |
| CN111079687A (en) * | 2019-12-26 | 2020-04-28 | 京东数字科技控股有限公司 | Certificate camouflage identification method, device, equipment and storage medium |
| CN111695406A (en) * | 2020-04-23 | 2020-09-22 | 西安电子科技大学 | Face recognition anti-spoofing method, system and terminal based on infrared ray |
| CN111754441A (en) * | 2020-06-29 | 2020-10-09 | 国网甘肃省电力公司电力科学研究院 | Passive detection method for image copy-paste forgery |
| CN112257685A (en) * | 2020-12-08 | 2021-01-22 | 成都新希望金融信息有限公司 | Face copying recognition method and device, electronic equipment and storage medium |
| CN112580621A (en) * | 2020-12-24 | 2021-03-30 | 成都新希望金融信息有限公司 | Identity card copying and identifying method and device, electronic equipment and storage medium |
| CN112528998A (en) * | 2021-02-18 | 2021-03-19 | 成都新希望金融信息有限公司 | Certificate image processing method and device, electronic equipment and readable storage medium |
Non-Patent Citations (1)
| Title |
|---|
| 基于RGB图像的人脸活体检测技术研究;张克越;《中国优秀硕士学位论文全文数据库 信息科技辑》;20210515(第05期);I138-1351 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113033530A (en) | 2021-06-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107944339B (en) | Certificate verification and identity verification method and device | |
| WO2020098250A1 (en) | Character recognition method, server, and computer readable storage medium | |
| Kraetzer et al. | Modeling attacks on photo-ID documents and applying media forensics for the detection of facial morphing | |
| CN110766033B (en) | Image processing method, image processing device, electronic equipment and storage medium | |
| CN108337505B (en) | Information acquisition method and device | |
| CN112528998B (en) | Certificate image processing method and device, electronic equipment and readable storage medium | |
| US20180232605A1 (en) | Determining image forensics using gradient statistics at edges | |
| CN116416190A (en) | Flaw detection method and device, electronic equipment and storage medium | |
| CN114638777A (en) | Image defect detection method, device, electronic equipment and medium | |
| CN113033530B (en) | Certificate copying detection method and device, electronic equipment and readable storage medium | |
| CN109656800B (en) | Method and device for testing image recognition application, terminal and storage medium | |
| CN113158773B (en) | Training method and training device for living body detection model | |
| CN114422144A (en) | Method, system, equipment and storage medium for improving reliability of chain certificate of scene certificate block | |
| CN110414522A (en) | A kind of character identifying method and device | |
| JP6195252B2 (en) | Mobile device for financial document transactions | |
| CN113239339A (en) | Certificate shooting method and device, computer equipment and storage medium | |
| US8483449B2 (en) | Registration device, checking device, program, and data structure | |
| CN112541899B (en) | Incomplete detection method and device of certificate, electronic equipment and computer storage medium | |
| CN111626244B (en) | Image recognition method, device, electronic equipment and medium | |
| CN114417397A (en) | Behavior portrait construction method and device, storage medium and computer equipment | |
| CN111767845B (en) | Certificate identification method and device | |
| CN112597810A (en) | Identity document authentication method and system | |
| WO2021073576A1 (en) | Intelligent vehicle verification method and system, computer device, and storage medium | |
| CN111160093A (en) | Method, device, equipment and storage medium for identifying certificate photo camouflage in certificate | |
| CN111860486A (en) | Card identification method, device and equipment |
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 |