WO2021143088A1 - Procédé et appareil de vérification synchrone pour de multiples types de certificats, et dispositif informatique et support de stockage - Google Patents

Procédé et appareil de vérification synchrone pour de multiples types de certificats, et dispositif informatique et support de stockage Download PDF

Info

Publication number
WO2021143088A1
WO2021143088A1 PCT/CN2020/103394 CN2020103394W WO2021143088A1 WO 2021143088 A1 WO2021143088 A1 WO 2021143088A1 CN 2020103394 W CN2020103394 W CN 2020103394W WO 2021143088 A1 WO2021143088 A1 WO 2021143088A1
Authority
WO
WIPO (PCT)
Prior art keywords
certificate
image
detected
document
detection
Prior art date
Application number
PCT/CN2020/103394
Other languages
English (en)
Chinese (zh)
Inventor
顾佳页
王波
孙建波
叶松
Original Assignee
深圳壹账通智能科技有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 深圳壹账通智能科技有限公司 filed Critical 深圳壹账通智能科技有限公司
Publication of WO2021143088A1 publication Critical patent/WO2021143088A1/fr

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features

Definitions

  • This application relates to the field of artificial intelligence image detection, and in particular to a method, device, computer equipment, and storage medium for synchronous detection of multiple document types.
  • the present application provides a method, device, computer equipment, and storage medium for synchronous detection of multiple certificate types, which realize rapid synchronous detection of multiple certificate types, thereby improving detection efficiency and reducing costs.
  • a synchronous detection method for multiple certificate types including:
  • the checklist includes Multiple document types
  • the recognition result includes a list of certificate types of all the images to be detected, and the list of certificate types includes the certificate types identified from all the images to be detected;
  • a synchronous detection device for multiple certificate types including:
  • the receiving module is used to receive the certificate detection instruction and obtain the image file to be detected containing the detection number; wherein the image file to be detected includes the image file to be detected;
  • the determining module is configured to obtain a preset numbering rule, and determine the checklist of the image file to be detected and the certificate detection model corresponding to the checklist according to the detection number of the image file to be detected and the number rule;
  • the checklist includes multiple types of documents;
  • the recognition module is used to input all the image pieces to be detected into the document detection model, extract document features from all the image pieces to be detected through the document detection model, and obtain the document detection model to output according to the document characteristics
  • the recognition result includes a list of document types for all the image pieces to be detected, and the list of document types includes the type of document identified from all the image pieces to be detected;
  • the detection module is used to confirm that the certificate type of the image file to be detected is qualified when the certificate type included in the check list is completely consistent with the certificate type in the certificate type list, and to mark the image file to be detected It is the detected image file and stored in the database.
  • a computer device includes a memory, a processor, and computer-readable instructions that are stored in the memory and can run on the processor, and the processor implements the following steps when the processor executes the computer-readable instructions:
  • the checklist includes Multiple document types
  • the recognition result includes a list of certificate types of all the images to be detected, and the list of certificate types includes the certificate types identified from all the images to be detected;
  • One or more readable storage media storing computer readable instructions, when the computer readable instructions are executed by one or more processors, the one or more processors execute the following steps:
  • the checklist includes Multiple document types
  • the recognition result includes a list of certificate types of all the images to be detected, and the list of certificate types includes the certificate types identified from all the images to be detected;
  • the method, device, computer equipment, and storage medium for synchronous detection of multiple certificate types obtained the image file to be detected (including multiple image files to be detected) containing the detection number, and determine the location according to the detection number and numbering rules.
  • the document features are extracted from all the images to be detected, and the recognition results output by the document detection model according to the features of the documents are obtained.
  • the recognition results include a list of the document types of all the images to be detected (including from all the said images).
  • the certificate type identified in the image file to be detected when the certificate type included in the checklist is exactly the same as the certificate type in the certificate type list, confirm that the certificate type of the image file to be detected is qualified and mark it at the same time It is the detected image file and stored in the database.
  • a certificate detection model that only targets all the certificate types included in the checklist is formed, and a one-to-one identification method is formed.
  • This application is more targeted. Moreover, the training time of the document detection model is shorter, and the neural network structure of the document detection model is simpler and more accurate. At the same time, this application realizes rapid and synchronous detection of multiple document types, and different document types can be mixed in one image. , It is not necessary to confirm in advance that a type of document is in an image to be recognized, thereby improving the detection efficiency and reducing the cost.
  • FIG. 1 is a schematic diagram of an application environment of a method for synchronous detection of multiple certificate types in an embodiment of the present application
  • FIG. 2 is a flowchart of a method for synchronous detection of multiple certificate types in an embodiment of the present application
  • FIG. 3 is a flowchart of a method for synchronous detection of multiple certificate types in another embodiment of the present application.
  • step S30 is a flowchart of step S30 of the method for synchronous detection of multiple certificate types in an embodiment of the present application
  • FIG. 5 is a flowchart before step S304 of the method for synchronous detection of multiple certificate types in an embodiment of the present application
  • FIG. 6 is a flowchart after step S304 of the method for synchronous detection of multiple certificate types in an embodiment of the present application
  • FIG. 7 is a flowchart after step S40 of the method for synchronous detection of multiple certificate types in an embodiment of the present application.
  • FIG. 8 is a flowchart after step S50 of the method for synchronous detection of multiple certificate types in an embodiment of the present application.
  • Fig. 9 is a schematic block diagram of a device for synchronous detection of multiple certificate types in an embodiment of the present application.
  • Fig. 10 is a schematic diagram of a computer device in an embodiment of the present application.
  • the multi-certificate type synchronization detection method provided by this application can be applied in the application environment as shown in Fig. 1, in which the client (computer equipment) communicates with the server through the network.
  • the client includes, but is not limited to, various personal computers, notebook computers, smart phones, tablet computers, cameras, and portable wearable devices.
  • the server can be implemented as an independent server or a server cluster composed of multiple servers.
  • a method for synchronous detection of multiple certificate types is provided, and the technical solution mainly includes the following steps S10-S40:
  • S10 Receive a certificate detection instruction, and obtain a to-be-detected image file containing a detection number; wherein the to-be-detected image file includes the to-be-detected image file.
  • the document detection instruction is received to obtain the image file to be inspected
  • the document inspection instruction is an instruction to select the image file to be inspected to be inspected and to be triggered after confirmation.
  • the method of obtaining can be set as required.
  • the method of obtaining may be to obtain the image file to be detected through the certificate detection instruction (the certificate detection instruction includes the image file to be detected). ). Obtain the image file to be detected and so on according to the storage path of the image file to be detected included in the certificate instruction.
  • the to-be-detected image files are a plurality of to-be-detected image pieces that need to be detected and identified, and the to-be-detected image pieces include images of multiple certificate types, and the to-be-detected image files include the detection number, and the The detection number is a unique number assigned to the detection of multiple certificate types that determine the image file to be detected.
  • the component factors of the detection number can be set according to requirements, for example, the component factors of the detection number can be set by Composed of numbers or/and letters, the image file to be detected may be a folder, the image file to be detected is a single file under the folder, and the attribute of the folder contains the detection number of the image file to be detected
  • the folder name of the image file to be detected contains the detection number WJJAF201909100
  • the image to be detected can also be a file, and the file contains multiple pages of the image to be detected
  • the properties of the folder include the The detection number of the image file to be detected, for example, the file name of the image file to be detected contains the detection number WJAF201909110.
  • the numbering rule is a rule formulated according to the requirements to establish a corresponding relationship between the detection category and the numbering range.
  • the numbering rule defines: 1.
  • the numbering range is WJJAF201909001 to WJJAF201909999.
  • the corresponding detection category is September 2019 security inspection.
  • the number range is WJJBX201908001 to WJJBX201908999.
  • the corresponding test category is the insurance test category in August 2019.
  • the detection category of the image file to be detected is determined, that is, the detection of the image file to be detected is determined according to the degree of matching between the detection number and the numbering rule Category, such as:
  • the detection number of the image file to be detected is WJJAF201909100
  • its detection number WJJAF201909100 falls within the number range WJJAF201909001 to WJJAF201909999 corresponding to the type of security detection.
  • the detection category of the image file to be detected is security detection
  • the check list corresponding to the detection category is determined and recorded as the check list of the image file to be detected.
  • the certificate detection model corresponding to the detection category is determined and recorded as the check list corresponding to the detection category.
  • the certificate detection model corresponding to the checklist of the image file to be detected wherein one of the detection categories corresponds to one of the checklists, one of the detection categories corresponds to one of the certificate detection models, and the checklist is for all Describes the checklist of multiple types of document types for the image files to be detected.
  • the type of security inspection corresponds to a checklist of ID cards, social security card documents, and security certificates.
  • the document inspection model corresponding to the type of security inspection is determined.
  • the credential detection model corresponding to the checklist of the image file to be detected is a deep convolutional neural network model trained and trained for different checklists.
  • the checklist corresponds to one credential detection model, and there are many There are a variety of document detection models corresponding to different checklists.
  • the checklist includes multiple types of certificates, for example, the checklist includes ID certificates, bank card certificates, social security card certificates, etc., and the certificate detection model only identifies all certificate types included in the checklist.
  • the document detection model is highly pertinent, and only detects the document types included in the checklist, and the training time of the document detection model is short, and the accuracy of the document detection model is higher.
  • the identification result includes a list of certificate types of all the image pieces to be detected, and the list of certificate types includes all the types of documents identified in the image pieces to be detected.
  • the document detection model extracts all the document features in the image pieces to be detected, and the document detection model includes the deep convolution that has been trained A neural network model, where the document features include texture features extracted from the deep convolutional neural network model, and according to the document features, the document detection model outputs the recognition result of the image to be detected, and the recognition result includes A list of the certificate types of all the image files to be detected (the certificate types identified in all the image files to be detected).
  • the identification process of the document detection model is as follows. First, the document detection model performs gray-scale processing on all the image parts to be detected, generates gray-scale images of all the image parts to be detected, and adopts an edge detection method.
  • the image is input to the trained deep convolutional neural network model, the local binary pattern feature map is extracted through the deep convolutional neural network model, and the deep convolutional neural network model output is obtained
  • the identification result of the identification result represents the identification type of the image of the identification area; finally, all identification types of the identification area image are written into the identification type list of all the images to be detected.
  • the credential detection model can simultaneously detect the image files to be detected of multiple credential types, and credential images of different credential types can be mixed in the same image to be detected, so that the image to be detected does not need to be required There is only one type of certificate for the file, which improves the detection efficiency and reduces the cost.
  • all the images to be detected are input into the credential detection model, and all the images to be detected are processed by the credential detection model. Extract the document characteristics, and obtain the recognition result output by the document detection model according to the document characteristics; the recognition result includes a list of the document types of all the images to be inspected, and the list of document types includes all the documents to be inspected.
  • the types of documents identified in the image include:
  • S301 Obtain all the image components to be detected, perform grayscale processing on all the image components to be detected, and generate grayscale images of all the image components to be detected.
  • the image to be inspected includes an image of three RGB channels (red channel, green channel, and blue channel), that is, each pixel in the image to be inspected has three channel component values.
  • the R component value red channel component value
  • G component value green channel component value
  • B component value blue channel component value
  • each pixel in the image to be detected Perform gray-scale processing, obtain the gray-scale value of each pixel through the weighted average method, thereby generating the gray-scale image of the image to be detected.
  • the image to be detected in the three channels is transformed into one The grayscale image of the channel, and then only one channel is processed, reducing the processing of each channel separately.
  • S302 Recognize the grayscale images of all the image parts to be detected by an edge detection method, and extract a number of document area images in the grayscale images.
  • the edge detection method is to identify the pixels in the image whose gray value is significantly different from the gray value around the pixel, because the image to be detected contains all the documents.
  • the brightness on both sides of the edge is obviously different, and the difference in gray value on both sides of the edge of all the documents after gray-scale processing is more obvious, so the gray-scale image of the image to be detected is identified through the edge detection method
  • the pixels with obvious changes in gray value are then extracted by extracting the area formed by the pixels with obvious changes in gray value (that is, the area where each document is located), and mark it as the document area in the gray image image.
  • the Local Binary Patterns (LBP) method uses each pixel as a central pixel, and the gray value corresponding to the pixel is used as the threshold of the central pixel, and the adjacent regions are The gray value corresponding to the pixel point within is compared with the threshold value of the central pixel. If the gray value of the adjacent pixel point is greater than the threshold value of the central pixel, the position of the pixel point is marked as binary code 1, otherwise it is Binary code 0, which combines the binary codes corresponding to the pixels in all adjacent areas in a clockwise order into a set of binary numbers, and the binary value is the corresponding local binary pattern feature value (LBP value).
  • LBP value local binary pattern feature value
  • the adjacent area is a 3 ⁇ 3 area of 8 pixels (excluding the central pixel), an 8-bit binary number is generated, and the range of the feature value of the local binary mode is an integer value from 0 to 255.
  • the local binary mode method since the local binary mode method has the advantages of rotation invariance and gray level invariance, the local binary mode feature map has strong robustness to illumination. Through image analysis, there are some significant texture features in different document images. Therefore, the identification of the texture feature through the local binary pattern method to detect the type of document improves the reliability and accuracy of the identification.
  • the local binary pattern feature map corresponding to each of the document area images is input into the trained deep convolutional neural network model in the document detection model, so as to compare the local binary pattern features Image recognition processing, that is, the texture feature recognition of the local binary pattern feature map, the recognition result based on the texture feature statistics can be obtained, that is, the recognition result of the document area image is obtained, and the recognition result At the same time, it characterizes the document type of the image of the document area, such as ID card, bank card, social security card, and so on.
  • S305 Write the certificate types of all the images of the certificate area into the list of certificate types of all the images to be inspected.
  • the document types of all the images of the document area identified in the document type list.
  • the document types are identified as ID documents and social security card documents, then the identity documents and social security card documents Write it in the list of certificate types.
  • the edge detection method is used to identify the pixels with obvious gray value changes in the grayscale image, so as to quickly extract the regional images of all documents.
  • the local binary mode method has rotation invariance and grayscale. With the advantage of degree invariance, the identification of texture features through the local binary pattern method for document type detection improves the accuracy and reliability of recognition.
  • the deep convolution The neural network model extracts the texture feature of the regional image and outputs the recognition result of the regional image according to the texture feature.
  • the recognition result characterizes the certificate type of the regional image, it includes:
  • the training image samples are all associated with a certificate type label, and the training image samples are local binary pattern feature maps corresponding to the grayscale images of the training image samples.
  • the initial neural network model acquires all model parameters of the YOLO model, and determines all the model parameters as the initial parameters of the initial neural network model.
  • the transfer learning uses the parameters of existing training models in other fields to apply to tasks in this field, that is, the initial neural network model obtains YOLO (You Only Only) through transfer learning. Look once) all the model parameters of the model, and then determine all the model parameters as the initial parameters of the initial neural network model.
  • the initial neural network model includes the initial parameters, and the training image samples are input into the initial neural network model.
  • the training image sample is processed by the initial neural network model, and the texture feature in the training image sample is extracted, and the texture feature includes the wave pattern feature, the pattern feature, and the abnormal spot feature.
  • S3045 Obtain a recognition result output by the initial neural network model according to the texture feature, and determine a loss value according to the degree of matching between the recognition result and the certificate type label.
  • the identification type of the training image sample is performed through the initial neural network model to obtain the recognition result of the initial neural network model,
  • the identification result of the training image sample is compared with the certificate type label of the training image sample to determine the corresponding loss value, that is, the loss value is calculated by the loss function of the initial neural network model.
  • the preset convergence condition may be a condition that the value of the loss value is small and will not drop after 3000 calculations, that is, the value of the loss value is small and does not fall after 3000 calculations When it will drop again, stop training, and record the converged initial neural network model as the trained deep convolutional neural network model;
  • the preset convergence condition can also be that the loss value is less than a set threshold The condition is that when the loss value is less than the set threshold, the training is stopped, and the converged initial neural network model is recorded as the trained deep convolutional neural network model.
  • the input training image samples of the initial neural network model are grayscale images that have been grayscale processed, there is no need to perform complex image transformation processing on the image in the input layer, and only need to recognize texture features, so the initial neural network
  • the model has fast processing speed and small capacity, so it can be used in portable mobile terminals with small capacity.
  • the method includes:
  • the initial parameters of the iterative initial neural network model are continuously updated to continuously move closer to the accurate recognition result, so that the accuracy of the recognition result becomes higher and higher.
  • the initial parameters of the iterative initial neural network model are continuously updated to continuously move closer to the accurate recognition result, so that the accuracy of the recognition result becomes higher and higher.
  • the method includes:
  • S306 Determine a preset information area in the document area image according to the document type of the document area image.
  • the certificate area image is acquired, and the certificate type of the certificate area image is acquired, and the preset information area in the certificate area image is determined according to the certificate type, wherein the different The certificate type matches the different preset information areas.
  • the preset information area corresponding to the ID card is a rectangular area at the position of the ID card number
  • the preset information area corresponding to the bank card certificate is the rectangular area at the position of the bank card number.
  • S307 Cut a cropped image corresponding to each of the preset information areas from the image of the document area.
  • the preset information area in the document area image is intercepted, and the intercepted image is marked as the corresponding cropped image.
  • S308 Transform the cut image by using an inverse binarization method to generate an inverse binarized cut image corresponding to the image of the document area.
  • the inverse binarization method is to first perform gray-scale binarization processing on the cropped image to obtain a binarized image, and then perform segmentation processing on the binarized image by selecting an appropriate threshold, and finally Inverting the binarized image after the segmentation process, thereby obtaining the inverse binarized cropped image.
  • the certificate number recognition model is a trained and trained deep neural network model, wherein the training method and network structure of the certificate number recognition model can be set according to requirements.
  • the certificate number The training mode of the recognition model is a migration learning training mode
  • the network structure of the ID number recognition model is a VGG16 network structure.
  • the digital feature and letter feature of the de-binarized crop image are extracted by the document number recognition model, and the document number recognition model recognizes the document area image according to the digital feature and the letter feature So that the certificate number recognition model outputs the certificate information associated with the certificate type of the certificate area image.
  • this application uses the deep neural network model to recognize the numbers and letters in the image, which speeds up the recognition speed and reduces the requirements for image quality. Because the deep neural network model recognizes through the number and letter features, input fuzzy images The effect of accurate identification can also be achieved.
  • S310 Write the credential information and the credential type of the regional image in a credential type list of all the image files to be detected in association with each other.
  • the method further includes:
  • the method for confirming the detection failure of the image file to be detected can be set according to requirements, for example, a corresponding failure window pops up, and the failure window indicates the "XXX certificate type of the image file to be detected" "Error" and other words.
  • the checklist and the certificate type list compare the certificate types contained in the checklist with the certificate types in the certificate type list, and the certificate types contained in the checklist
  • the certificate type is completely consistent with the certificate type in the certificate type list
  • the file is marked as a detected image file and stored in the database, and the detection result is recorded, so that the detection result corresponding to the user can be retrieved in the subsequent user authentication process.
  • This application obtains the image file to be inspected (including multiple image files to be inspected) containing the inspection number, and determines the check list of the image file to be inspected and the certificate detection model corresponding to the check list according to the inspection number and numbering rules ,
  • the check list contains multiple types of documents, all the image pieces to be detected are input into the document detection model, and the document features are extracted from all the image pieces to be detected through the document detection model to obtain the document detection
  • the model outputs the recognition results based on the document features, and the recognition results include a list of document types for all the images to be inspected (including the document types identified from all the images to be inspected), in the check list
  • the included certificate type is completely consistent with the certificate type in the certificate type list, it is confirmed that the certificate type of the image file to be detected is qualified, and the image file to be detected is also marked as a detected image file and stored in the database.
  • a certificate detection model that only targets all the certificate types included in the checklist is formed, and a one-to-one identification method is formed.
  • This application is more targeted. Moreover, the training time of the document detection model is shorter, and the neural network structure of the document detection model is simpler and more accurate. At the same time, this application realizes rapid and synchronous detection of multiple document types, and different document types can be mixed in one image. , It is not necessary to confirm in advance that a type of document is in an image to be recognized, thereby improving the detection efficiency and reducing the cost.
  • the certificate type list further includes certificate information associated with each of the certificate types identified from all the image files to be detected;
  • the method further includes:
  • S50 Obtain information to be audited associated with the detection number; the information to be audited includes certificate verification parameters corresponding to each of the certificate types included in the checklist.
  • the information to be reviewed is the information associated with the detection number, and includes the certificate verification parameters corresponding to the certificate type, for example: ID card number corresponds to XXXXXXXXXXXXXXXXXXXXX (13-digit certificate number), and bank card certificate corresponds to XXXXXXXXXXXXXX ( 16-digit card number), etc., wherein the certificate type included in the information to be reviewed is consistent with the certificate type included in the checklist.
  • the certificate verification parameters and the certificate information of the same certificate type are obtained, and the certificate verification parameters of the same certificate type are compared with the certificate information, and the certificate verification parameters are compared with the certificate information.
  • the credential information matches, it is confirmed that the credential type corresponding to the credential verification parameter is successfully verified.
  • the document information associated with all the document types contained in the image file to be inspected is verified successfully, and it is confirmed that the image file to be inspected has passed the verification, such as in the case of bad assets.
  • the user’s identity is verified by providing image files of the user’s identity. After all the document types in the image files of all user identities have been verified successfully, the user’s identity verification is confirmed.
  • the method includes:
  • the certificate verification parameter does not match the certificate information, that is, the certificate verification parameter is not equal to the certificate information, as long as one of the certificate verification parameters is not equal to the certificate information, then It is confirmed that the review result of the image file to be detected is not passed, and a failure prompt message of failing the review is prompted.
  • the failure prompt information can be set according to requirements, for example, a prompt "XXXX document information does not match" pops up.
  • the multi-document type synchronous detection device includes a receiving module 11, a determining module 12, an identifying module 13 and a detecting module 14.
  • the detailed description of each functional module is as follows:
  • the receiving module 11 is configured to receive a certificate detection instruction, and obtain a to-be-detected image file containing a detection number; wherein the to-be-detected image file includes the to-be-detected image file;
  • the determining module 12 is configured to obtain a preset numbering rule, and determine a checklist of the image file to be detected and a certificate detection model corresponding to the checklist according to the detection number of the image file to be detected and the number rule;
  • the checklist includes multiple types of certificates;
  • the recognition module 13 is configured to input all the image pieces to be detected into the document detection model, extract the document features from all the image pieces to be detected through the document detection model, and obtain the document detection model according to the document characteristics Output recognition result;
  • the recognition result includes a list of certificate types of all the image files to be detected, and the list of certificate types includes the certificate types identified from all the image files to be detected;
  • the detection module 14 is used for confirming that the certificate type of the image file to be detected is qualified when the certificate type included in the check list is exactly the same as the certificate type in the certificate type list, and the image file to be detected Mark the detected image file and store it in the database.
  • the device for simultaneous detection of multiple certificate types further includes:
  • An obtaining module configured to obtain information to be audited associated with the detection number; the information to be audited includes certificate verification parameters corresponding to each of the certificate types included in the checklist;
  • the comparison module is configured to compare the certificate verification parameters of the same certificate type with the certificate information, and when the certificate verification parameters match the certificate information, confirm the certificate verification parameters corresponding to the certificate verification parameters.
  • the certificate type verification is successful;
  • the confirmation module is used for confirming that the image file to be inspected has passed the review when all the certificate types of the image file to be inspected are verified successfully.
  • the acquisition module further includes:
  • the identification module 13 includes:
  • the detection failure unit is configured to confirm that the detection of the to-be-detected image file fails when the check list is inconsistent with the certificate type list.
  • the identification module 13 further includes:
  • the first acquiring unit is configured to acquire all the image parts to be inspected, perform gray-scale processing on all the image parts to be inspected, and generate gray-scale images of all the image parts to be inspected;
  • the first extraction unit is used to identify the gray-scale images of all the image parts to be detected by the edge detection method, and extract a number of document area images in the gray-scale images;
  • a conversion unit configured to convert each of the document area images into a local binary pattern feature map corresponding to each of the document area images by using a local binary mode method
  • the recognition unit is configured to input the local binary pattern feature map corresponding to each of the document region images into the trained deep convolutional neural network model in the document detection model, and pass the deep convolutional neural network model.
  • the network model extracts the texture feature of the local binary pattern feature map, and obtains the recognition result output by the deep convolutional neural network model according to the texture feature, and the recognition result represents the document of the document area image Types of;
  • the successful detection unit is used to write the certificate types of all the images of the certificate area into the list of certificate types of all the images to be detected.
  • the identification unit includes:
  • the second acquiring unit is configured to acquire training image samples; wherein each of the training image samples is associated with a certificate type label;
  • the migration unit is used to obtain all the model parameters of the YOLO model by the initial neural network model through migration learning, and determine all the model parameters as the initial parameters of the initial neural network model;
  • a second extraction unit configured to extract texture features in the training image sample through the initial neural network model
  • a determining unit configured to obtain a recognition result output by the initial neural network model according to the texture feature, and determine a loss value according to the degree of matching between the recognition result and the certificate type label;
  • the training completion unit is configured to record the initial neural network model after convergence as the trained deep convolutional neural network model when the loss value reaches a preset convergence condition.
  • the identification unit further includes:
  • a determining subunit configured to determine a preset information area in the image of the image of the image of the image of the image of the image of the image of the image of the preset of the preset information area of the preset information area according to the type of the image of the image of the image of the image of the image of the image of the image of the image of the image of the preset information area according to the type of the image of the image of the image of the image of the image of the image of the image of the image of the image;
  • a cropping subunit configured to crop a cropped image corresponding to each of the preset information areas from the image of the document area
  • a transforming subunit configured to transform the cropped image by using an inverse binarization method to generate an inverse binarized cropped image corresponding to the image of the document area;
  • the recognition subunit is used to input the de-binarized cropped image into a credential number recognition model, and the credential number recognition model extracts the digital features of the de-binarized cropped image corresponding to the credential area image And the letter feature, and output the recognition result of the de-binarized cropped image corresponding to the document area image according to the number feature and the letter characteristic, and the recognition result characterizes the difference with the document area image Document information associated with the document type;
  • the credential information is associated with the credential type of the regional image and written into the credential type list of all the image files to be detected.
  • Each module in the above-mentioned multi-certificate type synchronous detection device can be implemented in whole or in part by software, hardware, and a combination thereof.
  • the above-mentioned modules may be embedded in the form of hardware or independent of the processor in the computer equipment, or may be stored in the memory of the computer equipment in the form of software, so that the processor can call and execute the operations corresponding to the above-mentioned modules.
  • a computer device is provided.
  • the computer device may be a server, and its internal structure diagram may be as shown in FIG. 10.
  • the computer equipment includes a processor, a memory, a network interface, and a database connected through a system bus.
  • the processor of the computer device is used to provide calculation and control capabilities.
  • the memory of the computer device includes computer readable instructions and internal memory.
  • the computer readable instructions are stored with an operating system, computer readable instructions and a database.
  • the internal memory provides an environment for the operation of the operating system and the computer-readable instructions in the computer-readable instructions.
  • the network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer-readable instructions are executed by the processor, a method for synchronous detection of multiple certificate types is realized.
  • the readable storage medium provided in this embodiment includes a non-volatile readable storage medium and a volatile readable storage medium.
  • a computer device including a memory, a processor, and computer-readable instructions stored in the memory and capable of running on the processor. Synchronous detection method of certificate type.
  • one or more readable storage media storing computer readable instructions are provided.
  • the readable storage media provided in this embodiment include non-volatile readable storage media and volatile readable storage. Medium; the readable storage medium stores computer readable instructions, and when the computer readable instructions are executed by one or more processors, the one or more processors implement the method for synchronous detection of multiple certificate types in the foregoing embodiment.
  • Non-volatile memory may include read only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory.
  • Volatile memory may include random access memory (RAM) or external cache memory.
  • RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous chain Channel (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • General Physics & Mathematics (AREA)
  • Artificial Intelligence (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • General Engineering & Computer Science (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Multimedia (AREA)
  • Image Analysis (AREA)

Abstract

L'invention concerne un procédé et un appareil de vérification synchrone pour de multiples types de certificats, ainsi qu'un dispositif informatique et un support de stockage, qui se rapportent au domaine de la vérification d'image par intelligence artificielle. Le procédé consiste : à recevoir une instruction de vérification de certificat, et à acquérir un fichier image à vérifier qui comprend un numéro de contrôle ; en fonction du numéro de vérification du fichier image à vérifier et d'une règle de numérotation prédéfinie, à déterminer une liste de vérification et un modèle de vérification de certificat correspondant à celui-ci, la liste de vérification comprenant de multiples types de certificat ; à entrer tous les fragments d'image à vérifier dans le modèle de vérification de certificat, et au moyen de l'extraction de caractéristiques de certificat, à acquérir un résultat de reconnaissance de sortie comprenant une liste de types de certificat de tous les fragments d'image à vérifier ; et lorsque les types de certificat inclus dans la liste de vérification sont complètement cohérents avec des types de certificat dans la liste de types de certificat, à déterminer que le type de certificat du fichier image à vérifier est vérifié de telle sorte que celui-ci est aux normes, et à marquer le fichier image à vérifier en tant que fichier image vérifié et à stocker celui-ci dans une base de données. Au moyen du procédé, une vérification rapide et synchrone de types de certificats multiples est réalisée, ce qui permet d'améliorer l'efficacité de contrôle et de réduire les coûts.
PCT/CN2020/103394 2020-01-19 2020-07-22 Procédé et appareil de vérification synchrone pour de multiples types de certificats, et dispositif informatique et support de stockage WO2021143088A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202010061431.4 2020-01-19
CN202010061431.4A CN111275102A (zh) 2020-01-19 2020-01-19 多证件类型同步检测方法、装置、计算机设备及存储介质

Publications (1)

Publication Number Publication Date
WO2021143088A1 true WO2021143088A1 (fr) 2021-07-22

Family

ID=71003364

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/103394 WO2021143088A1 (fr) 2020-01-19 2020-07-22 Procédé et appareil de vérification synchrone pour de multiples types de certificats, et dispositif informatique et support de stockage

Country Status (2)

Country Link
CN (1) CN111275102A (fr)
WO (1) WO2021143088A1 (fr)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116402070A (zh) * 2023-05-31 2023-07-07 中电长城(长沙)信息技术有限公司 一种用于多证件的sp服务实现方法及系统
CN117391068A (zh) * 2023-10-27 2024-01-12 中国人寿保险股份有限公司山东省分公司 一种基于rpa的寿险保全业务信息核验方法及系统

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111275102A (zh) * 2020-01-19 2020-06-12 深圳壹账通智能科技有限公司 多证件类型同步检测方法、装置、计算机设备及存储介质
CN112418278A (zh) * 2020-11-05 2021-02-26 中保车服科技服务股份有限公司 一种多类物体检测方法、终端设备及存储介质
CN113657398B (zh) * 2021-08-18 2023-02-07 北京百度网讯科技有限公司 图像识别方法和装置
CN113743327A (zh) * 2021-09-07 2021-12-03 中国工商银行股份有限公司 单据识别方法、单据核对方法、装置和设备

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110008942A (zh) * 2019-01-15 2019-07-12 阿里巴巴集团控股有限公司 证件验证方法、装置、服务器及存储介质
US20190253254A1 (en) * 2018-02-06 2019-08-15 NB Research LLC System and method for securing a resource
CN110569769A (zh) * 2019-08-29 2019-12-13 浙江大搜车软件技术有限公司 图像识别方法、装置、计算机设备和存储介质
CN111275102A (zh) * 2020-01-19 2020-06-12 深圳壹账通智能科技有限公司 多证件类型同步检测方法、装置、计算机设备及存储介质

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20190253254A1 (en) * 2018-02-06 2019-08-15 NB Research LLC System and method for securing a resource
CN110008942A (zh) * 2019-01-15 2019-07-12 阿里巴巴集团控股有限公司 证件验证方法、装置、服务器及存储介质
CN110569769A (zh) * 2019-08-29 2019-12-13 浙江大搜车软件技术有限公司 图像识别方法、装置、计算机设备和存储介质
CN111275102A (zh) * 2020-01-19 2020-06-12 深圳壹账通智能科技有限公司 多证件类型同步检测方法、装置、计算机设备及存储介质

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116402070A (zh) * 2023-05-31 2023-07-07 中电长城(长沙)信息技术有限公司 一种用于多证件的sp服务实现方法及系统
CN116402070B (zh) * 2023-05-31 2023-08-08 中电长城(长沙)信息技术有限公司 一种用于多证件的sp服务实现方法及系统
CN117391068A (zh) * 2023-10-27 2024-01-12 中国人寿保险股份有限公司山东省分公司 一种基于rpa的寿险保全业务信息核验方法及系统
CN117391068B (zh) * 2023-10-27 2024-04-05 中国人寿保险股份有限公司山东省分公司 一种基于rpa的寿险保全业务信息核验方法及系统

Also Published As

Publication number Publication date
CN111275102A (zh) 2020-06-12

Similar Documents

Publication Publication Date Title
WO2021143088A1 (fr) Procédé et appareil de vérification synchrone pour de multiples types de certificats, et dispositif informatique et support de stockage
US10706277B2 (en) Storing anonymized identifiers instead of personally identifiable information
CN110287971B (zh) 数据验证方法、装置、计算机设备及存储介质
US10217179B2 (en) System and method for classification and authentication of identification documents using a machine learning based convolutional neural network
WO2021027336A1 (fr) Procédé et appareil d'authentification basés sur un cachet et une signature et dispositif informatique
US11349666B2 (en) Electronically signing and distributing identification data as a service that provides proof of identity, integrity, validity and origin of data for non-repudiation and ID validation methods
WO2020073502A1 (fr) Procédé et appareil de vérification de factures, procédé et appareil de vérification de financement, et dispositif ainsi que support
CN110569341B (zh) 配置聊天机器人的方法、装置、计算机设备和存储介质
CN108256591B (zh) 用于输出信息的方法和装置
CN111626124B (zh) Ocr图像样本生成、印刷体验证方法、装置、设备及介质
CN111191568B (zh) 翻拍图像识别方法、装置、设备及介质
CN108053545B (zh) 证件验真方法和装置、服务器、存储介质
CN110503099B (zh) 基于深度学习的信息识别方法及相关设备
CN113111880B (zh) 证件图像校正方法、装置、电子设备及存储介质
CA2975030A1 (fr) Systeme et methode de detection de caracteristiques frauduleuses sur les documents d'authentification
CN111144372A (zh) 车辆检测方法、装置、计算机设备和存储介质
CN112232336A (zh) 一种证件识别方法、装置、设备及存储介质
Raigonda Signature Verification System Using SSIM In Image Processing
CN113806613B (zh) 训练图像集生成方法、装置、计算机设备及存储介质
CN114386013A (zh) 学籍自动认证方法、装置、计算机设备及存储介质
CN113704111A (zh) 页面自动化测试方法、装置、设备及存储介质
CN111860314B (zh) 基于图像识别的电子证照验证方法、装置和系统
CN112686847B (zh) 身份证图像拍摄质量评价方法、装置、计算机设备和介质
CN110751140A (zh) 字符批量识别方法、装置和计算机设备
CN112380957A (zh) 他项权证信息识别方法、装置、计算机设备和存储介质

Legal Events

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

Ref document number: 20914315

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 14.11.2022)

122 Ep: pct application non-entry in european phase

Ref document number: 20914315

Country of ref document: EP

Kind code of ref document: A1