Identity auditing system for open group chat

Abstract

The invention belongs to the technical field of group chat auditing, and particularly relates to an identity auditing system for open group chat. Comprising a user uploading and interaction module, a document identification and preprocessing module, an automatic authentication and comparison module, an administrator auditing module, an auditing rule engine module, a user identity file and data management module, a security and privacy protection module, a group chat platform interface module and a background management and monitoring module. According to the method, multi-layer algorithms such as image enhancement, text extraction, structured extraction, true and false detection, face comparison and template verification are connected in series to form a multi-dimensional cross validation chain, comprehensive judgment is performed from multiple aspects of image authenticity, certificate consistency and certificate holder and certificate matching degree, fraudulent behaviors such as identity card counterfeiting, P graph counterfeiting and substitute video counterfeiting are greatly reduced, and the method is suitable for large-scale popularization and application. And the audit result credibility is improved.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of group chat auditing, and particularly relates to an identity auditing system for an open group chat. BACKGROUND

[0002] With the rapid development of various online social platforms and open group chat scenarios, the user scale is exponentially increasing, and a large number of strangers interact in the group chat, which makes the platform face higher risks in safety management. The existing group chat platforms usually only rely on mobile phone registration, simple nickname or avatar to identify the user identity, lack effective verification mechanism for the real identity of the user, and lead to frequent occurrence of identity fraud, false documents, malicious registration and fraud activities in the open group chat environment. Although some platforms have introduced real-name authentication, they are mostly limited to static picture uploading or manual checking, and have low auditing efficiency and limited accuracy, and lack support for more complex identification requirements such as video proof, certificate verification and forgery detection. At the same time, the existing auditing process mostly relies on manual auditing, which is labor-intensive and subjective, and is prone to omissions, and lacks fine-grained permission control and log auditing mechanism, and is difficult to meet the privacy protection requirements for sensitive identity information. In addition, the identity auditing process and the group chat system usually lack linkage, and cannot dynamically determine the joining and permission management of the user according to the auditing result, resulting in low management efficiency. SUMMARY

[0003] In view of the above situation, in order to overcome the defects of the prior art, the application provides an identity auditing system for an open group chat.

[0004] The technical scheme adopted by the application is as follows: the application provides an identity auditing system for an open group chat, which comprises: A user uploading and interacting module for enabling the user to submit an identity card, a certificate and a proof video; A document recognition and preprocessing module for processing the image / video uploaded by the user for subsequent identification and forgery detection; An automatic forgery detection and comparison module for helping the administrator to reduce workload and improve accuracy; An administrator auditing module for providing a manual auditing interface and operation tools; An auditing rule engine module for configuring auditing logic and automatic judgment conditions; A user identity archive and data management module for storing the user identity history and the auditing result; A security and privacy protection module for sensitive information; A group chat platform interface module for linkage with an open group chat platform; A background management and monitoring module for management and health monitoring.

[0005] Further, the document recognition and preprocessing module comprises: Image cleaning and enhancement unit, denoising, brightness enhancement, sharpening, distortion correction, certificate contour detection and cutting; Text extraction unit, identify ID information, identify key fields of certificates, extract certificate text frames displayed in videos; Structured field extraction unit, parse OCR information into structured fields, for comparison and archiving time, number, issuing unit field extraction and format normalization.

[0006] Further, the content display and group chat module includes: Group chat interface unit for real-time message display and basic interaction functions; Group content open display unit, visitors can browse part of it; Message management unit for sending, editing, withdrawing, shielding, and reporting messages; Group theme tag unit for giving the group a theme tag; Pinning and essence unit for visitors to quickly access key information.

[0007] Further, the automatic authentication and comparison module includes: Image authenticity detection unit to detect tampering traces and determine whether it is from a real shot, and perform frame-by-frame detection on the video to identify forgeries; Face and certificate comparison unit to extract facial features and compare them with ID photos, and detect whether the face displayed in the video and the ID photo are consistent; Certificate template and layout matching unit to verify the format, font, seal, and QR code according to official templates, and analyze and verify the certificate barcode.

[0008] Further, the administrator review module includes: Review workstation unit to display all user-submitted materials, including video playback, image zoom, and comparison mode tools; Artificial authentication assistance tool unit to mark suspicious areas automatically detected by the system, provide text comparison frames and certificate template comparison functions, and quickly pass or reject, fill in reasons; Multi-level approval process unit to support initial review and different administrator permission control.

[0009] Further, the review rule engine module includes: Rule configuration unit to configure automatic pass conditions and automatic reject conditions; Condition decision unit to score each case according to the rules and generate review suggestions; Workflow control unit to control the distribution of review tasks according to priority and administrator load.

[0010] Further, the user identity profile and data management module comprises: a user profile management unit that stores all identity information that passes the review, records the type of certificate, the validity period, and the upload record; a review record and log unit that records the operation history of each review; a data retrieval and statistical unit that supports administrators to retrieve by username and certificate type, and to statistically review the pass rate, the rate of forgery, and the operation time.

[0011] Further, the security and privacy protection module comprises: a data encryption unit that stores the transmission layer TLS encryption of sensitive fields and desensitization; a permission and access control unit that divides the function permissions of different administrators to prevent unauthorized browsing of identity card content; an operation audit unit that records the time, personnel, and purpose of accessing sensitive data to prevent leaks and misoperations.

[0012] The present scheme also discloses an operation method of an open group chat identity review system, mainly comprising the following steps: Step A1: First, through the user upload and interaction module, guide the user to submit the identity card, certificate, and related proof video for identity review, provide prompts, format verification, and upload progress display during the submission process, and safely receive and temporarily store the original data for subsequent identification and processing; Step A2: After the user upload and interaction module receives the user uploaded image or video, the document recognition and preprocessing module automatically performs cleaning, enhancement, distortion correction, and cutting operations, then extracts the certificate text and structured fields, so that the content is in a unified format for subsequent forgery detection, comparison, and review; Step A3: After the document recognition and preprocessing module completes the preprocessing, it automatically detects the forgery of the submitted materials, including identifying whether there are tampering traces, checking whether the image is a real shot, analyzing the video authenticity, then comparing the user's face with the certificate photo, and verifying the credibility of the layout, font, seal, or two-dimensional code of the certificate using official template rules to generate a preliminary forgery detection result; Step A4: The automatically processed materials are pushed to the administrator review module by the automatic forgery detection and comparison module, the administrator can view the images, videos, and suspicious areas marked by the system through the review workstation, and use artificial forgery detection auxiliary tools for magnification, comparison, template comparison, and risk confirmation, then perform preliminary review or re-review according to the permissions, and give artificial decisions of passing, rejecting, or supplementing materials; Step A5: In the review process, the rule engine module scores and judges the case based on the preset rules, automatically gives review suggestions, and combines the workflow control unit to distribute the review tasks according to the priority and administrator load, realizing the cooperation of automatic and manual review; Step A6: When the review is completed, the user identity file and data management module writes the user's identity information, certificate structured field, review results and operation log into the user identity file and data management module to form a searchable historical file and support subsequent statistical analysis, risk backtracking and re-verification; Step A7: The security and privacy protection module encrypts and desensitizes sensitive data throughout the process, and ensures that only authorized personnel can access related content through permission control and operation audit. Finally, the system and group chat platform interface module are linked to determine whether to open the user's chat permission or display the necessary information according to the review results, and the background monitoring module continuously monitors and maintains the system running state.

[0013] The beneficial effects obtained by the application with the above structure are as follows: the application provides an identity review system for open group chat, which realizes the following beneficial effects: (1) The multi-layer algorithms such as image enhancement, text extraction, structured extraction, true-false detection, face comparison and template verification are connected in series to form a "multi-dimensional cross verification chain", which comprehensively judges from the aspects of image authenticity, certificate consistency, and matching degree of certificate holder and certificate, greatly reduces fraud behaviors such as fake ID cards, P pictures and substitute videos, and improves the credibility of the review results.

[0014] (2) It contains automatic identification, automatic comparison and rule engine scoring mechanism, which can automatically filter out obviously fake or obviously valid materials before manual intervention of the administrator, only push the suspicious and manual confirmation materials to the administrator, greatly reduce the repetitive work, and make the review efficiency improve several times.

[0015] (3) All records of identity files, review records, access logs, permission operations are recorded to form a complete audit chain, which is convenient for later traceability, law enforcement cooperation and dispute handling. Combined with the encryption, desensitization and permission control of the privacy protection module, it can meet the supervision requirements and significantly reduce the risk of privacy leakage. BRIEF DESCRIPTION OF DRAWINGS

[0016] Figure 1 A flowchart of an identity review system for open group chat is provided for the application.

[0017] The accompanying drawings are used to provide a further understanding of the application, and constitute a part of the specification, together with the embodiments of the application, to explain the application, and do not constitute a limitation on the application. DETAILED DESCRIPTION

[0018] With reference to the drawings, the technical solutions in the embodiments of the present application will be clearly and completely described in order to make the technical solutions in the embodiments of the present application apparent to those skilled in the art. Obviously, the described embodiments are only a part but not all of the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work fall into the scope of the present application.

[0019] Embodiment 1 Please refer to Figure 1 As shown in the figure, the embodiment is an identity auditing system for an open group chat, which comprises a user uploading and interacting module, a document recognition and preprocessing module, an automatic identification and comparison module, an administrator auditing module, an auditing rule engine module, a user identity archive and data management module, a security and privacy protection module, a group chat platform interface module, and a background management and monitoring module. The user uploading and interacting module is configured to guide the user to submit an ID card, a certificate, and a proof video. The document recognition and preprocessing module is configured to process the image / video uploaded by the user for subsequent identification and identification. The automatic identification and comparison module is configured to help the administrator reduce workload and improve accuracy. The administrator auditing module is configured to provide a manual auditing interface and an operation tool. The auditing rule engine module is configured to configure the auditing logic and automatically determine the conditions. The user identity archive and data management module is configured to store the user identity history and the auditing result. The security and privacy protection module is configured to protect sensitive information. The group chat platform interface module is configured to be linked with the open group chat platform. The background management and monitoring module is configured to manage and monitor the health.

[0020] Embodiment 2 Please refer to Figure 1 As shown in the figure, the embodiment is a use method of an identity auditing system for an open group chat, which comprises the following steps: Step A1: first, through the user uploading and interacting module, guide the user to submit an ID card, a certificate, and a proof video for identity auditing, provide prompts, format verification, and uploading progress display during the submission process, and safely receive and temporarily store the original data for subsequent identification and processing; Step A2: after the user uploading and interacting module receives the image or video uploaded by the user, the document recognition and preprocessing module automatically performs cleaning, enhancement, distortion correction, and cutting operation, then extracts the certificate text and structured field, so that the content is in a unified format for subsequent identification, comparison, and auditing; Step A3: After pre-processing is completed, the document recognition and pre-processing module automatically identifies the submitted materials for forgery, including identifying whether there are traces of tampering, checking whether the image is a real shot, analyzing the authenticity of the video, then comparing the user's face with the certificate photo, and verifying the layout, font, seal or QR code credibility of the certificate using official template rules to generate a preliminary forgery result; Step A4: The automatic forgery identification and comparison module pushes the automatically processed materials to the administrator review module. Administrators can view images, videos and suspicious areas marked by the system through the review workstation, and use artificial forgery assistance tools for zooming, comparison, template comparison and risk confirmation. Then, according to the permission, preliminary review or review is carried out, and the artificial decision of passing, rejecting or supplementing materials is given; Step A5: In the review process, the rule engine module scores and judges the case based on preset rules, automatically gives review suggestions, and combines the workflow control unit to distribute the review tasks according to priority and administrator load, realizing the cooperation of automation and artificial review; Step A6: After the review is completed, the user identity file and data management module writes the user's identity information, certificate structured field, review result and operation log into the user identity file and data management module, forms a searchable historical archive, and supports subsequent statistical analysis, risk backtracking and re-verification; Step A7: The security and privacy protection module encrypts and desensitizes sensitive data throughout the process, and ensures that only authorized personnel can access related content through permission control and operation audit. Finally, the system and group chat platform interface module are linked to decide whether to open the user's chat permission or display the necessary information according to the review result, and the background monitoring module continuously monitors and maintains the system running state.

[0021] The text extraction unit uses CRNN+CTC model: 1. Feature extraction stage: Input image , after convolution layer: ; Convolution layer calculation method: ; Output feature map As a sequence input.

[0022] 2. Sequence modeling stage Convolution features are cut into sequences by column ; Bi-directional : Forward: ; Backward: ; Final sequence features: .

[0023] 3. Character probability distribution prediction: For each time step : ; Softmax formula: 4. Text alignment and output: CTC is used for unaligned input / output sequences, the true text is , and the prediction is ; CTC loss: ; Where: ; is the mapping that converts the path containing blanks to the final text.

[0024] 5. CTC greedy decoding: ; Rules of CTC: Select the character with the maximum probability at each time step; Merge duplicate characters; Delete blanks.

[0025] The structured field extraction unit uses a sequence labeling model + a rule expression + a field template matching: 1. Sequence labeling algorithm formula The OCR text sequence is represented as: ; Encoding: LSTM: ; Where ; CRF labeling layer: CRF finds the optimal label sequence : ; Where: : Label transition matrix; : The score of the current character belonging to the label ; : Normalization factor; The final field label is: ; 2. Regular field parsing Identity card number field expression: ; Date field: 3. Template matching: ; For ensuring the field format is correct, e.g.: ; Where: : field extracted from OCR; : official certificate template field; : string or feature similarity; : field weight; If: ; The structured field is considered valid.

[0026] Image authenticity detection unit Image authenticity detection can be designed as a binary classification problem, training a convolutional neural network; Input image ; Feature extraction, each layer of convolution calculation: ; Where: ; Convolution; Convolution kernel; Activation function such as ReLU; Classification layer: ; Output probability: Binary classification using Sigmoid: ; Fake probability: .

[0027] Face and certificate comparison unit can use FaceNet, ArcFace, CosFace to calculate face vector, and use distance to judge whether it is the same person: 1. Feature vector extraction: ID photo: ; Selfie or face in video frame: ; Deep convolutional network (ArcFace ResNet100); 2. ArcFace loss: ArcFace increases the angular interval , enhancing the discrimination: ;​ wherein: : sample to correct class center angle; : scaling factor; : angle interval.

[0028] 3. Comparison formula Cosine similarity: ; Judge whether it is the same person: ; Set to 0.4-0.6.

[0029] The certificate template and layout matching unit adopts template matching, key point matching, and layout geometric constraint inspection: 1. Template matching: normalized cross-correlation Certificate image and official template : ; High correlation position indicates consistent layout; 2. Key point and geometric consistency Detect key points and match: ; Estimate affine / perspective matrix using RANSAC , if the number of matching inliers satisfies: ; The layout is determined to be real.

[0030] 3. Region layout check Calculate the position error for each field region : ; If all , the layout is correct.

[0031] Although embodiments of the present application have been shown and described, it is to be understood that various modifications, substitutions, replacements and changes can be made to these embodiments without departing from the principles and spirit of the present application, and the scope of the present application is defined by the appended claims and their equivalents.

[0032] The above describes the present application and its embodiments, which are not restrictive, and the drawings only show one of the embodiments of the present application, and the actual structure is not limited thereto. In general, if a person skilled in the art is inspired by it, without departing from the purpose of the present application, without creative design, similar structure and embodiments of the technical solution can be designed, which should belong to the protection scope of the present application.

Claims

1. An identity verification system for an open group chat, the system comprising: Comprise: User upload and interaction module: for users to submit ID cards, certificates and proof videos; Document recognition and preprocessing module: for processing user uploaded images / videos for subsequent recognition and authentication; Automatic authentication and comparison module: to help administrators reduce workload and improve accuracy; Administrator review module: to provide manual review interface and operation tools; Review rule engine module: to configure review logic and automatic judgment conditions; User identity archive and data management module: to store user identity history and review results; Security and privacy protection module: to protect sensitive information; Group chat platform interface module: to link with open group chat platform; Background management and monitoring module: for management and health monitoring. 2.The identity auditing system of an open group chat according to claim 1, wherein: The document recognition and preprocessing module comprises: Image cleaning and enhancement unit, denoising, brightness enhancement, sharpening, distortion correction, certificate contour detection and cutting; Text extraction unit, identify ID card information, identify key fields of certificates, extract certificate text frames displayed in videos; Structured field extraction unit, parse OCR information into structured fields for comparison and archive time, number, issuing unit field extraction and format normalization. 3.The identity auditing system of an open group chat according to claim 2, wherein: Content display and group chat module includes: Group chat interface unit, for real-time message display, basic interaction functions; Group content open display unit, visitors can browse part of it; Message management unit, for sending, editing, withdrawing, shielding and reporting messages; Group theme tag unit, for giving the group a theme tag; Top and essence unit, for visitors to quickly obtain key information.

4. The identity auditing system for an open group chat of claim 3, wherein: The automatic authentication and comparison module comprises: Image authenticity detection unit, detects whether there are tampering traces, judges whether it comes from real shooting, and performs frame-by-frame detection on videos to identify forgeries; Face and certificate comparison unit, extracts face features, compares them with ID card photos, and detects whether the face displayed in the video and the ID card photo are consistent; Certificate template and layout matching unit, checks version, font, seal and QR code according to official template, analyzes and verifies certificate barcodes.

5. The identity auditing system for an open group chat of claim 4, wherein: The administrator review module comprises: Review workstation unit, displays all user-submitted materials, including video playback, image zooming and comparison mode tools; Manual authentication assistance tool unit, marks suspicious areas detected by the system, provides text comparison box and certificate template comparison function, and provides quick pass or reject and reason filling; Multi-level approval process unit, supports initial review and re-review with different administrator permission control. 6.The identity auditing system of an open group chat according to claim 5, wherein: The review rule engine module comprises: Rule configuration unit, configure automatic pass conditions and automatic reject conditions; Condition decision unit, score each case according to rules and generate review suggestions; Workflow control unit, controls the distribution of review tasks according to priority and administrator load.

7. The identity auditing system for an open group chat of claim 6, wherein: The user identity archive and data management module comprises: User archive management unit, stores all passed identity information, records certificate type, validity period and upload record; Review record and log unit, records the operation history of each review; Data retrieval and statistics unit, supports administrator retrieval by username and certificate type, and statistics of review pass rate, forgery rate and operation time. 8.The identity auditing system of an open group chat of claim 7, wherein: The security and privacy protection module comprises: A data encryption unit for storing layer encryption transmission layer (TLS) encryption sensitive field desensitization; An authority and access control unit for dividing the function authority of different administrators to prevent unauthorized browsing of identity card content; An operation audit unit for recording the time, personnel, and purpose of accessing sensitive data to prevent leakage and misoperation.

9. An operation method of an identity verification system of an open group chat, characterized by; The operation of the identity verification system of the open group chat according to claim 8 mainly comprises the following steps: Step A1: First, the user uploads and interacts with the module, guides the user to submit an identity card, certificate, and related proof video for identity verification, provides prompts, format verification, and upload progress display during the submission process, and safely receives and temporarily stores the original data for subsequent identification and processing; Step A2: After the user uploads and interacts with the module receives the user uploaded image or video, the document recognition and preprocessing module automatically performs cleaning, enhancement, distortion correction, and cutting operation, then extracts the certificate text and structured field, so that the content is in a unified format for subsequent authentication, comparison, and verification; Step A3: After the document recognition and preprocessing module completes the preprocessing, it automatically authenticates the submitted materials, including identifying whether there are tampering traces, checking whether the image is a real shot, analyzing the authenticity of the video, then comparing the user's face with the certificate photo, and verifying the credibility of the layout, font, seal, or two-dimensional code of the certificate using official template rules to generate a preliminary authentication result; Step A4: The automatically processed materials are pushed to the administrator verification module by the automatic authentication and comparison module. The administrator can view the images, videos, and suspicious areas marked by the system through the verification workbench, and use artificial authentication auxiliary tools for magnification, comparison, template comparison, and risk confirmation, then perform preliminary review or review according to the permission, and give artificial decision of passing, rejection, or supplementary materials; Step A5: In the verification process, the rule engine module scores and judges the case based on preset rules, automatically gives verification suggestions, and distributes the verification tasks according to priority and administrator load in combination with the workflow control unit, realizing the cooperation of automation and manual verification; Step A6: After the verification is completed, the user identity archive and data management module writes the user's identity information, certificate structured field, verification result, and operation log into the user identity archive and data management module, forming a searchable historical archive, and supporting subsequent statistical analysis, risk backtracking, and re-verification; Step A7: The security and privacy protection module encrypts and desensitizes the sensitive data in the entire process, and ensures that only authorized personnel can access the relevant content through permission control and operation audit. Finally, the system and group chat platform interface module are linked to determine whether to open the user's chat permission or display the necessary information according to the verification result, and the background monitoring module continuously monitors and maintains the system running state.

Images