Certificate authentic identification method and device, computer equipment and storage medium

Abstract

The embodiment of the invention provides a certificate authentic identification method and device, computer equipment and a storage medium, and belongs to the technical field of artificial intelligence. The method comprises the following steps: acquiring an original certificate image of a target certificate; performing text recognition degree evaluation on the original certificate image to obtain text recognition degree evaluation data; wherein the text recognition degree evaluation data represents the recognizable degree of the text in the original certificate image; screening out a selected certificate image from the original certificate images according to the text recognition degree evaluation data; wherein the text in the selected certificate image can be identified; performing text recognition on the selected certificate image to obtain text recognition information; comparing the text identification information with preset certificate verification information to obtain text comparison information; wherein the certificate verification information is certificate information used for verifying the authenticity of the target certificate; and performing authentication processing on the target certificate according to the text comparison information to obtain certificate authentication information. According to the embodiment of the invention, the accuracy of certificate authentic identification can be improved.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to a document authentication method and device, a computer device, and a storage medium. Background Art

[0002] As the requirements for information security increase, it is necessary to authenticate the certificate in advance before registering the certificate information to prevent the wrong certificate information from being written into the system. In the related art, the authentication of certificate information has replaced manual authentication with automatic authentication, and the automatic authentication method requires obtaining the certificate image in advance, and then performing text recognition on the certificate image to authenticate the certificate based on the identified certificate text information. However, the certificate image acquisition process will be affected by some external factors, such as changes in lighting or poor shooting technology, resulting in low clarity of the certificate image. If the certificate image with low clarity is directly used for text information recognition, it will affect the accuracy of the certificate text information recognition, thereby affecting the accuracy of certificate authentication. Therefore, how to improve the accuracy of certificate authentication has become a technical problem that needs to be solved urgently. Summary of the invention

[0003] The main purpose of the embodiments of the present application is to provide a method and apparatus for document authentication, a computer device and a storage medium, so as to improve the accuracy of document authentication.

[0004] To achieve the above-mentioned purpose, a first aspect of an embodiment of the present application provides a method for verifying a certificate, the method comprising:

[0005] Obtaining the original document image of the target document;

[0006] Performing text recognition evaluation on the original certificate image to obtain text recognition evaluation data; wherein the text recognition evaluation data represents the degree of recognizability of the text in the original certificate image;

[0007] Filtering out a selected document image from the original document image according to the text recognition evaluation data; wherein the text in the selected document image can be recognized;

[0008] Performing text recognition on the selected document image to obtain text recognition information;

[0009] Comparing the text recognition information with preset certificate verification information to obtain text comparison information; wherein the certificate verification information is certificate information used to verify the authenticity of the target certificate;

[0010] The target certificate is authenticated according to the text comparison information to obtain certificate authentication information.

[0011] In some embodiments, performing text recognition evaluation on the original document image to obtain text recognition evaluation data includes:

[0012] Performing edge detection processing on the original document image to obtain image edge detection information;

[0013] Performing cropping processing on the original document image according to the image edge detection information to obtain a candidate document image;

[0014] Performing region cropping on the candidate document image according to preset text region information to obtain a target text region image; wherein the preset text region information indicates a text region for authenticity verification in the candidate document image;

[0015] A text recognition degree evaluation is performed on the target text area image to obtain the text recognition degree evaluation data.

[0016] In some embodiments, performing text recognition on the selected document image to obtain text recognition information includes:

[0017] Performing text positioning on the selected document image to obtain text positioning information;

[0018] Performing denoising processing on the selected document image according to the text positioning information to obtain a target document image;

[0019] The text recognition information is obtained according to the text positioning information, the preset text area information and the text recognition performed on the target document image.

[0020] In some embodiments, performing text positioning on the selected document image to obtain text positioning information includes:

[0021] Extracting features from the selected document image to obtain document image features;

[0022] Performing enhancement processing on the document image features to obtain enhanced image features;

[0023] Performing text region detection on the selected document image according to the enhanced image features to obtain text region detection information;

[0024] The selected document image is subjected to text positioning processing according to the text area detection information to obtain the text positioning information.

[0025] In some embodiments, performing denoising on the selected document image according to the text positioning information to obtain a target document image includes:

[0026] Performing Gaussian noise implantation on the selected document image according to the text positioning information to obtain random noise data;

[0027] The selected document image is denoised according to the random noise data to obtain a target document image.

[0028] In some embodiments, the obtaining the text recognition information according to the text positioning information and the preset text area information and the performing text recognition on the target document image includes:

[0029] Extracting features of the target document image according to the text positioning information and the preset text area information to obtain image text features;

[0030] Performing feature interaction on the image text features to obtain feature interaction information;

[0031] Annotating the image text features according to the feature interaction information to obtain feature annotation information;

[0032] Text recognition is performed on the target document image according to the feature annotation information to obtain the text recognition information.

[0033] In some embodiments, the step of selecting a selected document image from the original document image according to the text recognition evaluation data includes:

[0034] Comparing the text recognition evaluation data with a preset recognition threshold to obtain recognition comparison information;

[0035] The selected document image is screened out from the original document images according to the recognition comparison information.

[0036] To achieve the above-mentioned purpose, a second aspect of the embodiments of the present application provides a document authentication device, the device comprising:

[0037] An image acquisition module, used to acquire an original document image of a target document;

[0038] A text recognition evaluation module, used to perform text recognition evaluation on the original document image to obtain text recognition evaluation data; wherein the text recognition evaluation data represents the degree of recognition of the text in the original document image;

[0039] An image screening module, used for screening selected document images from the original document images according to the text recognition evaluation data; wherein the text in the selected document images can be recognized;

[0040] A text recognition module, used to perform text recognition on the selected document image to obtain text recognition information;

[0041] A text comparison module, used to compare the text recognition information with preset certificate verification information to obtain text comparison information; wherein the certificate verification information is certificate information used to verify the authenticity of the target certificate;

[0042] The authentication module is used to perform authentication processing on the target certificate according to the text comparison information to obtain certificate authentication information.

[0043] To achieve the above-mentioned purpose, the third aspect of an embodiment of the present application proposes a computer device, which includes a memory and a processor, the memory stores a computer program, and the processor implements the method described in the first aspect when executing the computer program.

[0044] To achieve the above-mentioned purpose, the fourth aspect of an embodiment of the present application proposes a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method described in the first aspect is implemented.

[0045] The document authentication method and device, computer equipment and storage medium proposed in the present application evaluate the text recognition degree of each original document image to obtain text recognition evaluation data before performing document authentication, and select document images with recognizable text from multiple original document images according to the evaluated text recognition evaluation data, so as to screen out the original document images with unrecognizable text content and leave the selected document images with recognizable text content. Therefore, text recognition is performed on the selected document images to obtain more accurate text recognition information. Then, the text recognition information and the document verification information are compared to obtain text comparison information, and the target document is authenticated according to the text comparison information, which can improve the authentication accuracy of the target document. Therefore, the document images with recognizable text are screened out to participate in the authentication operation of the target document, so as to screen out the document images with unrecognizable text due to environmental factors, so as to improve the recognition accuracy of the text in the document image, and then the document authentication operation is performed by using the text recognition information with higher recognition accuracy, which can improve the accuracy of the document authentication operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 is a flow chart of a method for verifying a certificate provided by an embodiment of the present application;

[0047] Figure 2 is a schematic diagram of an identity document in the document authentication method provided in an embodiment of the present application;

[0048] Figure 3 yes Figure 1 Flow chart of step S102 in FIG.

[0049] Figure 4It is a schematic diagram of the anti-tampering content in the original certificate image in the certificate authentication method provided in the embodiment of the present application;

[0050] Figure 5 It is a structural diagram of the ShuffleNetV2 model in the document authentication method provided in the embodiment of the present application;

[0051] Figure 6 yes Figure 1 Flow chart of step S103 in FIG.

[0052] Figure 7 yes Figure 1 Flow chart of step S104 in FIG.

[0053] Figure 8 yes Figure 7 Flow chart of step S701 in FIG.

[0054] Fig. 9 is a schematic diagram of a selected document image in the document authentication method provided in an embodiment of the present application;

[0055] Fig.10 yes Figure 7 Flowchart of step S702 in FIG.

[0056] Fig.11 It is a schematic diagram of the change process of the selected certificate image to the transparent window area in the target certificate image in the certificate authentication method provided in the embodiment of the present application;

[0057] Fig.12 yes Figure 7 Flow chart of step S703 in FIG.

[0058] Fig.13 is a schematic diagram of target certificate authentication in the certificate authentication method provided in an embodiment of the present application;

[0059] Fig.14 It is a structural schematic diagram of a certificate authentication device provided in an embodiment of the present application;

[0060] Fig.15 It is a schematic diagram of the hardware structure of the computer device provided in the embodiment of the present application. DETAILED DESCRIPTION

[0061] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0062] It should be noted that, although the functional modules are divided in the device schematic diagram and the logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than the module division in the device or the order in the flowchart. The terms "first", "second", etc. in the specification, claims and the above drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence.

[0063] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which this application belongs. The terms used herein are only for the purpose of describing the embodiments of this application and are not intended to limit this application.

[0064] First, some nouns involved in this application are analyzed:

[0065] Artificial intelligence (AI) is a new technical science that studies and develops theories, methods, technologies and application systems for simulating, extending and expanding human intelligence. AI is a branch of computer science that attempts to understand the essence of intelligence and produce a new type of intelligent machine that can respond in a similar way to human intelligence. Research in this field includes robotics, speech recognition, image recognition, natural language processing and expert systems. AI can simulate the information process of human consciousness and thinking. AI is also a theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0066] Image Forensics: refers to the detection of whether an image has been tampered or tampered. The purpose of the detection is to determine the authenticity and reliability of the image. Among them, the forensic technology of image tampering mainly analyzes the characteristics of digital images to identify the authenticity, integrity and source of the image. In other words, the forensic technology of image tampering mainly determines whether the content of the image is authentic after it is generated by the imaging device, whether the image has been tampered with, and from which device it is generated.

[0067] ShuffleNet: is a lightweight convolutional neural network model optimized for mobile and embedded devices. This model includes ShuffleNetV1, ShuffleNetV2, ShuffleNetV2+, ShuffleNetV2Large, ShuffleNetV2ExLarge, etc. ShuffleNetV2 improves some deficiencies in the V1 version by introducing the Channel Split operation, such as reducing the use of group convolutions, reducing MAC, and improving the accuracy of the network through feature reuse. While maintaining a large number of channels, ShuffleNetV2 avoids the use of dense convolutions and large group convolutions, achieving a high-efficiency and high-capacity model design.

[0068] Differentiable Binarization (DB): is an advanced technique for scene text detection. It integrates the binarization process directly into the segmentation network to achieve adaptive threshold prediction, thereby simplifying the post-processing process and improving detection performance. The core of the differentiable binarization algorithm lies in the differentiable binarization strategy, which allows the binarization process to be optimized while training the segmentation network. This is achieved by introducing an approximate binarization function, which calculates an approximate binary map based on the probability map and the threshold map, making the entire process differentiable and easy to train end-to-end.

[0069] Diffusion Models (DM): A generative model that has made significant progress in the field of artificial intelligence in recent years, especially in computer vision and natural language processing. Its core idea is to generate new samples by simulating the diffusion and denoising process of data, which includes the forward diffusion process (noise addition process) and the reverse diffusion process (denoising process).

[0070] Convolutional Recurrent Neural Network (CRNN): is a sequence-to-sequence network structure that combines convolutional neural network (CNN) and recurrent neural network (RNN) to process text data with sequence characteristics. CNN is used to capture spatial features in images, while RNN processes sequence information and recognizes text by capturing temporal dependencies. By combining these two networks, CRNN can process spatial and temporal information in images at the same time, thus performing well in text recognition tasks.

[0071] Image tampering detection mainly identifies and locates the modified area in the digital image to determine whether the digital image has been tampered with. In order to improve the security of information entry, image tampering detection technology will be integrated into the information entry system for tampering detection. For example, in the document authentication scenario, the target document is photographed and content recognized by the information entry system to obtain the document information, and then the document information is entered for authentication. However, during the document information shooting process, it will be affected by lighting changes or external environmental factors, resulting in low quality of the document image, which affects the accuracy of subsequent document authentication.

[0072] Based on this, the embodiment of the present application provides a document authentication method and device, a computer device and a storage medium, which is intended to collect the original document image of the target document, first perform text recognition evaluation on each original document image to obtain text recognition evaluation data that characterizes the degree of text recognition in the original document image. Selected document images whose text content can be recognized are screened out from the original document images according to the text recognition evaluation data, and then the text in the selected document image is recognized, and the recognized text is compared with the verification text of the target document to realize the authentication of the target document, so that the accuracy of the authentication of the target document is improved.

[0073] The certificate authentication method and apparatus, computer equipment and storage medium provided in the embodiments of the present application are specifically described through the following embodiments. First, the certificate authentication method in the embodiments of the present application is described.

[0074] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Among them, artificial intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0075] AI basic technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, mechatronics, etc. AI software technologies mainly include computer vision technology, robotics technology, biometrics technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0076] The certificate authentication method provided in the embodiment of the present application relates to the field of artificial intelligence technology. The certificate authentication method provided in the embodiment of the present application can be applied to a terminal, can also be applied to a server side, and can also be software running in a terminal or a server side. In some embodiments, the terminal can be a smart phone, a tablet computer, a laptop computer, a desktop computer, etc.; the server side can be configured as an independent physical server, or a server cluster or a distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms; the software can be an application that implements the certificate authentication method, etc., but is not limited to the above forms.

[0077] The present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer computer devices, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, etc. The present application can be described in the general context of computer executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application can also be practiced in distributed computing environments, in which tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0078] It should be noted that in each specific implementation of the present application, when it comes to the need to perform relevant processing based on data related to user identity or characteristics such as user information, user behavior data, user historical data, and user location information, the user's permission or consent will be obtained first, and the collection, use, and processing of these data will comply with relevant laws, regulations, and standards. In addition, when the embodiment of the present application needs to obtain the user's sensitive personal information, the user's separate permission or consent will be obtained through a pop-up window or by jumping to a confirmation page. After clearly obtaining the user's separate permission or consent, the necessary user-related data for the normal operation of the embodiment of the present application will be obtained.

[0079] Figure 1 is an optional flow chart of the certificate authentication method provided in the embodiment of the present application. Figure 1 The method may include but is not limited to steps S101 to S106.

[0080] Step S101, obtaining an original document image of a target document;

[0081] Step S102, performing text recognition evaluation on the original document image to obtain text recognition evaluation data; wherein the text recognition evaluation data represents the degree of recognition of the text in the original document image;

[0082] Step S103, selecting a selected document image from the original document images according to the text recognition evaluation data; wherein the text in the selected document image can be recognized;

[0083] Step S104, performing text recognition on the selected document image to obtain text recognition information;

[0084] Step S105, comparing the text recognition information with preset certificate verification information to obtain text comparison information; wherein the certificate verification information is certificate information used to verify the authenticity of the target certificate;

[0085] Step S106, performing authentication processing on the target certificate according to the text comparison information to obtain certificate authentication information.

[0086] In steps S101 to S106 shown in the embodiment of the present application, text recognition evaluation data is obtained by evaluating the text recognition of the original certificate image of the target certificate, and the text recognition evaluation data represents the degree of text recognition in the original certificate image, so as to screen out the selected certificate image that can be used for text recognition from the original certificate image according to the text recognition evaluation data. Therefore, by selecting the certificate image for text recognition, the recognized text is compared with the verification text of the target certificate, and the target certificate is authenticated based on the text comparison information. Therefore, by screening out the certificate images with identifiable text to participate in the authentication operation of the target certificate, the certificate images with unrecognizable text due to environmental factors are screened out, and the recognition accuracy of the text in the certificate image is improved. Then, by performing the authentication operation of the certificate through the text recognition information with higher recognition accuracy, the accuracy of the certificate authentication operation can be improved.

[0087] In step S101 of some embodiments, the target certificate is a certificate with tamper-proof content, which is content that is not easily tampered by common technology. It should be noted that the tamper-proof content is set on the original certificate through image processing, digital watermarking and encryption algorithm to obtain the target certificate.

[0088] For example, Figure 2As shown, if in the field of insurance services, it is necessary to enter the identity information of the insured user, the target document is an identity document, and a transparent window area T is set on the identity document. The transparent window area T is composed of a transparent material, and numbers and letters are printed in the transparent window area T as tamper-proof content. Due to the particularity of the transparent material, the transparent window area T is not easy to forge, so the text recognized by the transparent window area T can be used as the text for the target document authentication to determine whether the target document has been tampered with, thereby realizing the authentication of the target document.

[0089] See also Figure 3 In some embodiments, step S102 may include but is not limited to steps S301 to S304:

[0090] Step S301, performing edge detection processing on the original document image to obtain image edge detection information;

[0091] Step S302, cropping the original document image according to the image edge detection information to obtain a candidate document image;

[0092] Step S303, performing region cropping on the candidate document image according to preset text region information to obtain a target text region image; wherein the preset text region information indicates a text region for authenticity verification in the candidate document image;

[0093] Step S304: perform text recognition evaluation on the target text area image to obtain text recognition evaluation data.

[0094] In step S301 of some embodiments, edge detection is to identify the pixels with drastic brightness changes in the original document image to extract the contour and structural information of the object, so the sudden changes of the brightness or grayscale value of the pixels in the original document image are analyzed by edge detection, and the edge area in the original document image is determined according to the pixel brightness or grayscale value to obtain the image edge detection information. Specifically, the pixel brightness value is obtained by performing pixel brightness recognition on the original document image, and the area with a pixel brightness value greater than a preset brightness threshold is defined as the document edge area to obtain the image edge detection information.

[0095] In step S302 of some embodiments, the original document image is cropped according to the image edge detection information, that is, the area outside the edge of the original document image is cropped to obtain an image containing only the area corresponding to the target document as the candidate document image.

[0096] In step S303 of some embodiments, the preset text area information is used to indicate the preset text area for authenticity verification in the candidate certificate image, and is used to implement the preset text area for authentication of the target certificate. Specifically, the preset text area is the area corresponding to the anti-tampering content, so that the candidate certificate image is regionally cropped according to the preset text area, so that the area corresponding to the anti-tampering content is used as the target text area image. It should be noted that the target text area image is an image that only contains the anti-tampering content area, so as to reduce the interference of other texts in the candidate certificate image, and there is no need to perform text recognition evaluation on all the texts in the original certificate image, so that the text recognition evaluation of the original certificate image is more efficient.

[0097] In step S304 of some embodiments, the target text area image is evaluated for text recognition, that is, the degree of recognizability of the text in the target text area image is identified to obtain text recognition evaluation data, so as to determine whether the tamper-proof content in the original document image can be recognized based on the text recognition evaluation data. Specifically, this embodiment uses the ShuffleNet model to perform text recognition evaluation on the candidate document image. The ShuffleNet model first extracts the target text area of ​​the tamper-proof content in the candidate document image, and then classifies the target text area to obtain a text quality classification result. It should be noted that the text quality classification result includes two text quality evaluation values, which are defined as F[0] and F[1], respectively. F[0] represents the probability of good quality of the tamper-proof content in the target text area, and F[1] represents the probability of poor quality of the tamper-proof content in the target text area. In this embodiment, any one of the text quality evaluation values ​​is used as text recognition evaluation data.

[0098] For example, Figure 4 As shown, Figure 4 The tamper-proof content on the left is not clear, Figure 4 The tamper-proof content on the right is clearly visible. Figure 4 The document image on the left is used for authentication of the target document. Recognition errors are prone to occur during the text recognition process, so it is necessary to use a document image with clearly visible text for authentication of the target document, which can improve the accuracy of document authentication.

[0099] Specifically, this embodiment uses the ShuffleNet model to evaluate the text recognition of the candidate document image, and the ShuffleNet model in this embodiment is specifically the ShuffleNetV2 model. The structure of the ShuffleNetV2 model is as follows: Figure 5 As shown, through Figure 5It can be seen that at the beginning of each unit, the candidate document image is subjected to feature extraction to obtain candidate image features, and the candidate image features are divided into two branches, one of which does not change the candidate image features, and the other branch consists of three convolutional layers, through which the candidate feature image is convolved, and the middle convolutional layer is a 3*3 convolutional layer. Finally, the candidate image features after convolution processing are concatenated and input into the channel shuffle layer to enhance the communication between the candidate image features of the two branches through the channel shuffle layer to output a fused image feature with a higher degree of fusion, and finally the fused image features are input into the fully connected layer to output text recognition evaluation data that more accurately represents the degree of text recognition. Among them, the present embodiment sets two fully connected layers to obtain more accurate text recognition evaluation data.

[0100] In steps S301 to S305 shown in this embodiment, the area outside the edge of the original document image is first cropped to obtain a candidate document image that only contains the corresponding area of ​​the target document, and then the text area for authenticity verification in the candidate document image is extracted as the target text area image, and then the target text area image is evaluated for text recognition. This can not only improve the efficiency of text recognition evaluation, but also reduce the interference of other texts in the text recognition evaluation process, thereby improving the accuracy of text recognition evaluation.

[0101] See also Figure 6 In some embodiments, step S103 may include but is not limited to steps S601 to S602:

[0102] Step S601, comparing the text recognition evaluation data with a preset recognition threshold to obtain recognition comparison information;

[0103] Step S602: Select the document image from the original document images according to the recognition comparison information.

[0104] In step S601 of some embodiments, the recognition threshold can be customized or determined based on the text recognition accuracy of the selected document image. No specific restrictions are imposed on the recognition threshold here. Specifically, the selected document image is screened out based on the initial threshold, the text recognition information and text verification information of the selected document image are used to determine the recognition accuracy, and the initial threshold is adjusted based on the recognition accuracy to obtain the recognition threshold. As disclosed above, the text recognition evaluation data can be any text quality evaluation value, and the recognition comparison information is obtained by comparing the text recognition evaluation data with the recognition threshold, that is, comparing the text quality evaluation value with the recognition threshold, so as to determine whether the tamper-proof content in the original document image is clearly visible through the recognition comparison information.

[0105] In step S602 of some embodiments, if the text quality evaluation value represents the probability that the tamper-proof content in the target text area has poor quality, the recognition comparison information represents the original document image whose text quality evaluation value is less than or equal to the recognition threshold as the selected document image. If the text quality evaluation value represents the probability that the tamper-proof content in the target text area has good quality, the recognition comparison information represents the original document image whose text quality evaluation value is greater than or equal to the recognition threshold as the selected document image. Therefore, the selected document image whose tamper-proof content is clearly visible is selected from the original document image.

[0106] It should be noted that if the selected document image cannot be screened out from the original document image, that is, there is no original document image with clearly visible anti-tampering content, a prompt message is generated, and the prompt message is "The current document image cannot be recognized, please retake it." Therefore, if the anti-tampering content in the original document image is not clear, the prompt message prompts the user to retake the image of the target document, so that the authentication of the target document can be completed smoothly and accurately.

[0107] In steps S601 to S602 shown in the present embodiment, by comparing the text recognition evaluation data with the recognition threshold, selected document images in which the text can be recognized are screened out from the original document images based on the recognition comparison information. Then, the subsequent authentication of the target document based on the selected document images can improve the accuracy of the authentication judgment.

[0108] See also Figure 7 In some embodiments, step S104 may include but is not limited to steps S701 to S703:

[0109] Step S701, performing text positioning on the selected document image to obtain text positioning information;

[0110] Step S702, performing denoising processing on the selected document image according to the text positioning information to obtain a target document image;

[0111] Step S703, performing text recognition on the target document image according to the text positioning information and the preset text area information to obtain text recognition information.

[0112] In step S701 of some embodiments, text positioning is to determine the position of the text, letters, and numbers in the selected document image, so as to combine the positions of the text, letters, and numbers in the selected document image into text positioning information. It should be noted that there are multiple text positioning information on the selected document image, but the authentication of the target document only requires the text corresponding to the anti-tampering content, so it is necessary to find the area corresponding to the anti-tampering content according to the text positioning information, so as to authenticate the target document according to the identification text of the anti-tampering content.

[0113] In step S702 of some embodiments, in order to further improve the authenticity of the tamper-proof content in the selected document image and make the tamper-proof content identification more accurate, the selected document image is denoised according to the text positioning information to restore the text in the selected document image so that the identified text identification information is more accurate.

[0114] In step S703 of some embodiments, text recognition is performed on the target document image according to the text positioning information and the preset text area information, that is, the text area in the target document image whose text positioning information is the preset text area information is recognized to obtain text recognition information.

[0115] In steps S701 to S703 shown in this embodiment, by determining the text positioning information in the selected document image, and performing denoising on the selected document image according to the text positioning information, a target text image with text closer to the actual situation is obtained. Therefore, text recognition is performed on the target document image according to the text positioning information and the preset text area to obtain more accurate text recognition information.

[0116] See also Figure 8 In some embodiments, step S701 may include but is not limited to steps S801 to S804:

[0117] Step S801, extracting features from the selected document image to obtain document image features;

[0118] Step S802, enhancing the features of the document image to obtain enhanced image features;

[0119] Step S803, performing text region detection on the selected document image according to the enhanced image features to obtain text region detection information;

[0120] Step S804: Perform text positioning processing on the selected document image according to the text area detection information to obtain text positioning information.

[0121] It should be noted that in order to achieve accurate text positioning, this embodiment uses a segmentation-based text detection model to locate text on the selected document image, specifically a differentiable binarization model, referred to as a DB text detection model. The DB text detection model introduces a differentiable binarization module into the segmentation model, so that the DB text detection model can be binarized through an adaptive threshold map, and the adaptive threshold map can calculate the loss, which can not only improve the effect of text positioning, but also simplify the post-processing process. Specifically, the DB text detection model includes: an image feature extraction network, a feature enhancement network, and a text area probability prediction network.

[0122] In step S801 of some embodiments, feature extraction is performed on the selected document image through an image feature extraction network to obtain document image features. Specifically, the image feature extraction network of this embodiment includes a first convolution layer and a downsampling layer, and convolution processing is performed on the selected document image through the first convolution layer, and then downsampling is performed through the downsampling layer to output the document image features.

[0123] In step S802 of some embodiments, in order to enhance the features of the document image, the present embodiment performs enhancement processing on the features of the document image through a feature enhancement network to obtain enhanced image features. Specifically, the feature enhancement network is a convolutional network that efficiently extracts features of the document image features in various dimensions as enhanced image features through convolution operations.

[0124] In step S803 of some embodiments, the present embodiment performs text probability prediction on the enhanced image features through the text region probability prediction network to obtain text prediction probability, and locates text region detection information from the selected document image according to the text prediction probability and the preset probability threshold. Specifically, the enhanced image features whose text prediction probability is greater than or equal to the preset probability threshold are used as text features, and the region corresponding to the text features in the selected document image is used as text region detection information, so that the text positioning operation is accurate.

[0125] In step S804 of some embodiments, the text area detection information is the area where the text is located, and text positioning processing is performed on the selected document image based on the text area detection information, that is, the area where the text is located is marked to obtain text positioning information with markings.

[0126] Specifically, after the selected document image is input into the DB text detection model, the DB text detection model outputs text area detection information, and the text area in the selected document image is annotated according to the text area detection information to obtain the following Fig. 9 The selected document image is shown.

[0127] In steps S801 to S804 shown in this embodiment, the document image features in the selected document image are extracted and enhanced into enhanced image features, and then text area detection is performed on the selected document image based on the enhanced image features to obtain accurate text area detection information. Then, text positioning is performed on the selected document image based on the text area detection information, so that more accurate text positioning can be achieved.

[0128] See also Fig.10 In some embodiments, step S702 includes but is not limited to steps S1001 to S1002:

[0129] Step S1001, Gaussian noise is implanted into the selected document image according to the text positioning information to obtain random noise data;

[0130] Step S1002, denoising the selected document image according to the random noise data to obtain a target document image.

[0131] It should be noted that this embodiment uses a diffusion model to perform denoising on the selected document image, specifically a stable diffusion generation model, which can generate an image with more realistic text from an image with text. Specifically, the stable diffusion generation model includes two processing processes, specifically a forward process and a reverse process. The forward process is also called a diffusion process, and the diffusion process is a process of gradually adding Gaussian noise to the data until the data is programmed with random noise, and the reverse process is a process of decoding the random noise and restoring it to real data.

[0132] In step S1001 of some embodiments, a text region is located in the selected document image according to the text location information, and then Gaussian noise is gradually added to the text region in the selected document image to convert the selected document image into random noise data. Specifically, the stable diffusion generation model of this embodiment includes an encoder and a decoder, and the selected document image is diffused by the encoder and the text location information to output random noise data.

[0133] In step S1002 of some embodiments, the selected document image is denoised based on the random noise data, that is, the random noise data is restored to a target document image closer to the actual situation through a decoder, and then the target document image replaces the selected document image.

[0134] For example, Fig.11 , Fig.11 The figure shows the process of the selected document image changing from the diffusion model to the transparent window area in the target document image. Fig.11 It can be seen that after the selected document image is processed by the diffusion model, the output text is closer to the target document image of the real scene. Therefore, performing text recognition on the target document image can improve the accuracy of text recognition.

[0135] In steps S1001 to S1002 shown in this embodiment, the selected document image is converted into random noise data according to the text positioning information, and the random noise data is then restored to the target document image, so as to replace the selected document image with the target document image to achieve denoising of the selected document image. Then, the text in the target document image is closer to the real text of the target document, and text recognition is performed on the target document image, which can improve the accuracy of text recognition.

[0136] See also Fig.12In some embodiments, step S703 may include but is not limited to steps S1201 to S1204:

[0137] Step S1201, extracting features of the target document image according to the text positioning information and the preset text area information to obtain image text features;

[0138] Step S1202, performing feature interaction on the image text features to obtain feature interaction information;

[0139] Step S1203, annotating the image text features according to the feature interaction information to obtain feature annotation information;

[0140] Step S1204: Perform text recognition on the target document image according to the feature annotation information to obtain text recognition information.

[0141] In step S1201 of some embodiments, in order to improve the efficiency of text recognition and reduce text recognition in unnecessary areas, feature extraction is performed on the target document image based on the text positioning information and the preset text area information. Specifically, the selected text area is located in the target document image based on the text positioning information and the preset text area information, and the features of the corresponding area are extracted from the target document image based on the selected text area to obtain the image text features. It should be noted that the text recognition of the target document image in this embodiment is implemented using a convolutional recurrent neural model, which can recognize the text in the area corresponding to the preset text area information in the target document image. Among them, the convolutional recurrent neural model includes: a second convolutional layer, a recurrent layer and a transcription layer, and the image text features are obtained by extracting features of the target document image through the second convolutional layer, text positioning information and the preset text area information.

[0142] For example, if the target document is an identity document, the image text feature is the text feature corresponding to the transparent window area.

[0143] In step S1202 of some embodiments, as disclosed above, the convolutional recurrent neural network includes a recurrent layer. In this embodiment, the recurrent layer uses a bidirectional recurrent neural network to predict image text features, and learns each image text feature to achieve feature interaction between image text features and output feature interaction information. It should be noted that feature interaction information is information output by predicting and learning each image text feature, so it is more accurate to identify the text in the target document image through feature interaction information.

[0144] In step S1203 of some embodiments, the image text features are annotated through the recurrent layer and feature interaction information to obtain feature annotation information. It should be noted that the feature annotation information represents the probability distribution information corresponding to each annotated text, and the text content in the area corresponding to the preset text area information in the target document image can be clearly understood through the feature annotation information.

[0145] In step S1204 of some embodiments, text recognition is performed on the target document image according to the feature annotation information, that is, the feature annotation information is converted through the transcription layer to obtain the target annotation information, and then text recognition is performed on the target document image according to the target annotation information to obtain text recognition information. It should be noted that by converting the feature annotation information into the final annotation sequence as the target annotation information, and the sum of the probabilities corresponding to the annotation texts in the target annotation information is 1, the identification information of the anti-tampering content in the target document image can be more clearly understood through the target annotation information.

[0146] In steps S1201 to S1204 shown in this embodiment, the accuracy of text recognition can be improved by extracting image text features corresponding to preset text area information in the target document image, performing feature interaction and annotation processing on the image text features to obtain feature annotation information, and determining text recognition information of the area corresponding to the preset text area information in the target document image based on the feature annotation information.

[0147] In step S105 of some embodiments, the certificate verification information is certificate information used to verify the authenticity of the target certificate, and the certificate verification information can be extracted from the original certificate image or preset. Specifically, if the target certificate is an identity document, the text recognition information is the recognition information of the anti-tampering content, and the identity document will also be provided with text corresponding to the anti-tampering content, so the text corresponding to the anti-tampering content is the certificate verification information, and the text comparison information is obtained by comparing the text recognition information and the certificate verification information, and the authenticity of the target certificate can be determined by the text comparison information. Specifically, if the text comparison information is characterized by the text recognition information and the certificate verification information being the same, it is determined that the target certificate is a forged certificate; if the text comparison information is characterized by the text recognition information and the certificate verification information being different, it is determined that the target certificate is a genuine certificate.

[0148] In step S106 of some embodiments, the target certificate is authenticated according to the text comparison information, that is, the certificate authentication information is determined according to the text comparison information. Specifically, if the text comparison information indicates that the text recognition information and the certificate verification information are the same, the certificate authentication information is false; if the text comparison information indicates that the text recognition information and the certificate verification information are different, the certificate authentication information is determined to be true. The authenticity of the target certificate can be directly determined by the certificate authentication information.

[0149] Specifically, for example Fig.13 As shown, if the recognition text information of area A of the target certificate is Z68123, and the certificate verification information extracted from area B of the original certificate image corresponding to the target certificate is Z68132, then the text recognition information and the certificate verification information are inconsistent, and the target certificate is judged to be false, so as to achieve rapid authentication processing of the target certificate.

[0150] The embodiment of the present application is explained by taking the identification of identity documents as an example. The identity document is photographed to obtain the original document image, and then the edge detection of the original document image is performed to obtain the image edge detection information. According to the image edge detection information, the original document image is cropped to obtain the candidate document image containing only the corresponding area of ​​the identity document. In the process of identifying the identity document, the area corresponding to the anti-tampering content on the identity document needs to be judged whether it can be recognized, so the area corresponding to the candidate document image is cropped according to the preset text area information to obtain the target text area image containing only the area corresponding to the anti-tampering content. In this embodiment, the ShuffleNetV2 model is used to evaluate the text recognition of the target text area image. Specifically, the convolution layer in the ShuffleNetV2 model is used to extract the candidate image features from the target text area image, and then the candidate image features are fused into the fused image features, and then the fused image features are normalized by the fully connected layer to obtain the text recognition evaluation data. Among them, the text recognition evaluation data is the text quality evaluation value, and the text quality evaluation value can represent the probability of good or poor quality of the anti-tampering content in the target text area. If the text quality evaluation value is represented by the probability that the anti-tampering content in the target text area has poor quality, the original document image with a text quality evaluation value less than or equal to the recognition threshold is used as the selected document image; if the text quality evaluation value is represented by the probability that the anti-tampering content in the target text area has good quality, the original document image with a text quality evaluation value greater than or equal to the recognition threshold is used as the selected document image. Therefore, according to the recognition degree of the target text area in the original document image, a selected document image with a good recognition degree is selected, and then the target document is more accurately authenticated by the selected document image. The selected document image is positioned by the DB text detection model, specifically, the selected document image is extracted by the image feature extraction network to obtain the document image feature, the document image feature is enhanced by the feature enhancement network to obtain the enhanced image feature, and then the enhanced image feature is predicted by the text area probability prediction network to obtain the text prediction probability, and the text area detection information is located from the selected document image according to the text prediction probability and the preset probability threshold, and the selected document image is positioned by the text area detection information to obtain the text positioning information with annotations. Then, the DB text detection model outputs the text positioning information to the stable diffusion generation model, and the stable diffusion generation model includes an encoder and a decoder. The selected ID image is diffused through the encoder and the text positioning information to output random noise data, and then the random noise data is restored to a target ID image that is closer to the actual situation through the decoder, and the target ID image replaces the selected ID image.The stable diffusion generation model inputs the output target ID image into the convolutional recurrent neural model, wherein the convolutional recurrent neural model includes a second convolutional layer, a recurrent layer, and a transcription layer; the target ID image is feature extracted through the second convolutional layer, text location information, and preset text area information to obtain image text features, and the image text features are interactively learned through the recurrent layer, and annotated to obtain feature annotation information, so as to determine the identification information corresponding to the anti-tampering content in the target ID image according to the feature annotation information to obtain text recognition information. Finally, the text recognition information and the ID verification information are compared. If the text recognition information and the ID verification information are the same, the ID is determined to be a real ID; if the text recognition information and the ID verification information are different, the ID is determined to be a forged ID, so as to achieve more accurate identification of IDs and improve the identification capability of IDs.

[0151] See also Fig.14 The present application also provides a device for verifying a certificate, which can implement the above-mentioned method for verifying a certificate. The device includes:

[0152] The image acquisition module 1401 is used to acquire the original document image of the target document;

[0153] The text recognition evaluation module 1402 is used to perform text recognition evaluation on the original document image to obtain text recognition evaluation data; wherein the text recognition evaluation data represents the degree of recognition of the text in the original document image;

[0154] The image screening module 1403 is used to screen out a selected document image from the original document image according to the text recognition evaluation data; wherein the text in the selected document image can be recognized;

[0155] The text recognition module 1404 is used to perform text recognition on the selected document image to obtain text recognition information;

[0156] The text comparison module 1405 is used to compare the text recognition information with the preset certificate verification information to obtain text comparison information; wherein the certificate verification information is certificate information used to verify the authenticity of the target certificate;

[0157] The authentication module 1406 is used to perform authentication processing on the target certificate according to the text comparison information to obtain certificate authentication information.

[0158] The specific implementation of the certificate authentication device is substantially the same as the specific implementation of the certificate authentication method described above, and will not be described in detail herein.

[0159] The embodiment of the present application also provides a computer device, the computer device includes a memory and a processor, the memory stores a computer program, and the processor implements the above-mentioned document authentication method when executing the computer program. The computer device can be any intelligent terminal including a tablet computer, a car computer, etc.

[0160] See also Fig.15 , Fig.15 The hardware structure of a computer device according to another embodiment is shown, and the computer device includes:

[0161] The processor 1501 may be implemented by a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (Application Specific Integrated Circuit, ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of the present application;

[0162] The memory 1502 can be implemented in the form of a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 1502 can store an operating system and other application programs. When the technical solution provided in the embodiment of this specification is implemented by software or firmware, the relevant program code is stored in the memory 1502, and the processor 1501 calls and executes the document authentication method of the embodiment of this application;

[0163] Input / output interface 1503, used to implement information input and output;

[0164] Communication interface 1504, used to realize communication interaction between the device and other devices, which can be realized through wired mode (such as USB, network cable, etc.) or wireless mode (such as mobile network, WIFI, Bluetooth, etc.);

[0165] A bus 1505 that transmits information between the various components of the device (e.g., the processor 1501, the memory 1502, the input / output interface 1503, and the communication interface 1504);

[0166] The processor 1501 , the memory 1502 , the input / output interface 1503 and the communication interface 1504 are connected to each other in communication within the device via the bus 1505 .

[0167] The embodiment of the present application also provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the above-mentioned document authentication method is implemented.

[0168] The memory, as a non-transient computer-readable storage medium, can be used to store non-transient software programs and non-transient computer executable programs. In addition, the memory may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some embodiments, the memory may optionally include a memory remotely disposed relative to the processor, and these remote memories may be connected to the processor via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0169] The document authentication method and device, computer equipment and storage medium provided in the embodiments of the present application first evaluate the text recognition of the original document image of the target document before the document authentication, and extract the selected document image with text recognition from the original document image based on the text recognition evaluation data obtained by the evaluation, and then identify the text recognition information of the selected document image. Finally, the target document is authenticated based on the text comparison information between the text recognition information and the document verification information to improve the accuracy of the document authentication.

[0170] The embodiments described in the embodiments of the present application are intended to more clearly illustrate the technical solutions of the embodiments of the present application and do not constitute a limitation on the technical solutions provided in the embodiments of the present application. Those skilled in the art will appreciate that with the evolution of technology and the emergence of new application scenarios, the technical solutions provided in the embodiments of the present application are also applicable to similar technical problems.

[0171] Those skilled in the art will appreciate that the technical solutions shown in the figures do not constitute a limitation on the embodiments of the present application, and may include more or fewer steps than shown in the figures, or a combination of certain steps, or different steps.

[0172] The device embodiments described above are merely illustrative, and the units described as separate components may or may not be physically separated, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0173] Those skilled in the art will appreciate that all or some of the steps in the methods disclosed above, and the functional modules / units in the systems and devices may be implemented as software, firmware, hardware, or a suitable combination thereof.

[0174] The terms "first", "second", "third", "fourth", etc. (if any) in the specification of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0175] It should be understood that in the present application, "at least one (item)" means one or more, and "plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" can mean: only A exists, only B exists, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally indicates that the objects associated before and after are in an "or" relationship. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0176] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the above units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0177] The units described above as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0178] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0179] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including multiple instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of various embodiments of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (Read-Only Memory, referred to as ROM), random access memory (Random Access Memory, referred to as RAM), disk or optical disk and other media that can store programs.

[0180] The preferred embodiments of the present invention are described above with reference to the accompanying drawings, but the scope of the rights of the present invention is not limited thereto. Any modification, equivalent substitution and improvement made by a person skilled in the art without departing from the scope and essence of the present invention should be within the scope of the rights of the present invention.

Claims

1. A method for verifying a certificate, characterized in that: The method comprises: Obtaining the original document image of the target document; Performing text recognition evaluation on the original certificate image to obtain text recognition evaluation data; wherein the text recognition evaluation data represents the degree of recognizability of the text in the original certificate image; Filtering out a selected document image from the original document image according to the text recognition evaluation data; wherein the text in the selected document image can be recognized; Performing text recognition on the selected document image to obtain text recognition information; Comparing the text recognition information with preset certificate verification information to obtain text comparison information; wherein the certificate verification information is certificate information used to verify the authenticity of the target certificate; The target certificate is authenticated according to the text comparison information to obtain certificate authentication information.

2. The method according to claim 1, characterized in that The step of performing text recognition evaluation on the original document image to obtain text recognition evaluation data includes: Performing edge detection processing on the original document image to obtain image edge detection information; Performing cropping processing on the original document image according to the image edge detection information to obtain a candidate document image; Performing region cropping on the candidate document image according to preset text region information to obtain a target text region image; wherein the preset text region information indicates a text region for authenticity verification in the candidate document image; A text recognition degree evaluation is performed on the target text area image to obtain the text recognition degree evaluation data.

3. The method according to claim 2, characterized in that The performing text recognition on the selected document image to obtain text recognition information includes: Performing text positioning on the selected document image to obtain text positioning information; Performing denoising processing on the selected document image according to the text positioning information to obtain a target document image; The text recognition information is obtained according to the text positioning information, the preset text area information and the text recognition performed on the target document image.

4. The method according to claim 3, characterized in that The performing text positioning on the selected document image to obtain text positioning information includes: Extracting features from the selected document image to obtain document image features; Performing enhancement processing on the document image features to obtain enhanced image features; Performing text region detection on the selected document image according to the enhanced image features to obtain text region detection information; The selected document image is subjected to text positioning processing according to the text area detection information to obtain the text positioning information.

5. The method according to claim 3, characterized in that: The step of performing denoising on the selected document image according to the text positioning information to obtain a target document image includes: Performing Gaussian noise implantation on the selected document image according to the text positioning information to obtain random noise data; The selected document image is denoised according to the random noise data to obtain a target document image.

6. The method according to claim 3, characterized in that: The obtaining of the text recognition information according to the text positioning information and the preset text area information and the performing of text recognition on the target document image includes: Extracting features of the target document image according to the text positioning information and the preset text area information to obtain image text features; Performing feature interaction on the image text features to obtain feature interaction information; Annotating the image text features according to the feature interaction information to obtain feature annotation information; Text recognition is performed on the target document image according to the feature annotation information to obtain the text recognition information.

7. The method according to any one of claims 1 to 6, characterized in that: The step of selecting a selected document image from the original document image according to the text recognition evaluation data comprises: Comparing the text recognition evaluation data with a preset recognition threshold to obtain recognition comparison information; The selected document image is screened out from the original document images according to the recognition comparison information.

8. A document authentication device, characterized in that: The device comprises: An image acquisition module, used to acquire an original document image of a target document; A text recognition evaluation module, used to perform text recognition evaluation on the original document image to obtain text recognition evaluation data; wherein the text recognition evaluation data represents the degree of recognition of the text in the original document image; An image screening module, used for screening selected document images from the original document images according to the text recognition evaluation data; wherein the text in the selected document images can be recognized; A text recognition module, used to perform text recognition on the selected document image to obtain text recognition information; A text comparison module, used to compare the text recognition information with preset certificate verification information to obtain text comparison information; wherein the certificate verification information is certificate information used to verify the authenticity of the target certificate; The authentication module is used to perform authentication processing on the target certificate according to the text comparison information to obtain certificate authentication information.

9. A computer device, characterized in that: The computer device comprises a memory and a processor, the memory stores a computer program, and the processor implements the document authentication method according to any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the document authentication method according to any one of claims 1 to 7 is implemented.

Images