Insurance slip anti-counterfeiting method, device and equipment based on two-dimensional code and storage medium

Abstract

The invention relates to an insurance slip anti-counterfeiting method based on a two-dimensional code, and the method comprises the steps: obtaining insurance slip data information which comprises an insurance slip number, a product code, an insurance applicant name and an insurance applicant identity certificate number; obtaining a unique identity identifier of a business operator, and performing MD5 hash operation on the identifier to generate a signature string; the generated signature string is used as an encryption key, the insurance slip data information is used as an encryption original text, and encryption operation is carried out to generate an encrypted data string; and converting the generated encrypted data string into two-dimensional code graph data, and embedding a two-dimensional code graph into the insurance slip to form the insurance slip of the anti-counterfeiting two-dimensional code graph. By encrypting the insurance policy information and generating the two-dimensional code, the insurance policy content is effectively prevented from being tampered and counterfeited, the security and authenticity of data are ensured, the anti-counterfeiting performance of the insurance policy information is improved, the workload of manual auditing is reduced, the auditing efficiency is improved, and the risk in the business process is reduced.

Description

Technical Field

[0001] The present invention relates to the fields of data security technology and financial technology, and in particular to an insurance policy anti-counterfeiting method, device, equipment and storage medium based on a QR code. Background Art

[0002] In group property insurance business, the development of electronic processes is relatively lagging, and paper documents still account for a high proportion of the business process. The diversity and non-standardized formats of paper documents make it difficult to implement automated audit technology. The current audit process mainly relies on manual operations, which not only has a large workload and low efficiency, but also has limited audit coverage, making it difficult to ensure the authenticity of all insurance policies, increasing the risk of missed inspections and misjudgments in operations. This situation not only affects the efficiency of business processing, but also brings greater business compliance risks.

[0003] Although the existing OCR (optical character recognition) technology has certain applications in certain standardized scenarios, the actual application effect of OCR is limited in group property insurance due to the wide variety of products and ever-changing policy formats. Existing technology is difficult to conduct effective image recognition training for thousands of product categories, which greatly reduces the production value of this technology in policy review and cannot cover non-standardized documents such as homemade policies. These technical limitations make it difficult to automate policy review, and the current situation of relying on manual review cannot be effectively improved.

[0004] Furthermore, insufficient verification of the authenticity of the underwriting information, especially the insurance application, in group property insurance directly leads to potential exposure to claims risk. Once an insurance application with false information enters the underwriting process, the insurance company will face serious legal and economic risks in the claims process. If the insurance company cannot provide effective underwriting basis due to the authenticity of the insurance application, it may face a situation of losing the case, resulting in huge economic losses. In addition, the limitations of manual review will also delay the speed of claims decision-making, affect customer experience, and further increase the complexity and uncertainty of business processing.

[0005] With the development trend of digitalization and intelligence in the insurance industry, how to solve the problems of low efficiency of document review and insufficient authenticity verification in group property insurance business has become the key to breakthroughs in the industry. Summary of the invention

[0006] The main purpose of the present invention is to provide an anti-counterfeiting method, device, equipment and storage medium for insurance applications based on a QR code, aiming to solve the technical problem that the prior art cannot effectively prevent the insurance application information from being tampered with or forged and cannot fully guarantee the authenticity of the insurance application.

[0007] To achieve the above object, the present invention provides an insurance policy anti-counterfeiting method based on a QR code, comprising:

[0008] Obtaining insurance application data information, the insurance application data information including the insurance application number, product code, policyholder name and policyholder identification number;

[0009] Obtain a unique identity identifier of the service operator, perform an MD5 hash operation on the unique identity identifier, and generate a signature string;

[0010] The signature string is used as an encryption key, and the insurance application data information is used as the encrypted original text to perform an encryption operation to generate an encrypted data string;

[0011] The encrypted data string is converted into two-dimensional code graphic data, and the two-dimensional code graphic data is embedded into the insurance application form to generate an insurance application form with an anti-counterfeiting two-dimensional code graphic.

[0012] In one embodiment, after the insurance application form with the anti-counterfeiting two-dimensional code graphic is generated, the following steps are further included:

[0013] Obtaining an insurance application form to be verified, and extracting a data string to be verified from the anti-counterfeiting two-dimensional code graphic of the insurance application form to be verified;

[0014] Decrypting the data string to be verified to restore the data information to be verified;

[0015] Identify the insurance application number from the data information to be verified, and confirm the corresponding original insurance application data information based on the insurance application number;

[0016] Compare the data information to be verified with the data information of the original insurance application form to verify whether the contents are consistent;

[0017] If the contents are consistent, a pre-trained deep learning model is used to identify whether the target seal exists in the insurance application to be verified;

[0018] If the target seal is identified, it is confirmed that the insurance application to be verified has passed the verification.

[0019] In one embodiment, identifying whether a target seal exists in the insurance application to be verified by a pre-trained deep learning model includes:

[0020] Extracting a seal image from the insurance application form to be verified, removing noise and performing binarization processing on the seal image to separate the seal area from the background;

[0021] Inputting the processed seal image into a pre-trained deep learning model to identify the seal type in the seal image, wherein the seal type includes a company seal and a business-specific seal;

[0022] According to the recognized seal type, compare it with the type of the target seal;

[0023] If the seal type does not match, confirm that the target seal does not exist in the insurance application to be verified;

[0024] If the seal type matches, extracting feature information corresponding to the identified seal type from the seal image, the feature information including the shape and texture of the seal;

[0025] Matching the characteristic information with the characteristics of the target seal, and analyzing the similarity between the seal image and the target seal;

[0026] If the similarity exceeds the similarity threshold, it is confirmed that the target seal exists in the insurance application to be verified.

[0027] In one embodiment, after comparing the data information to be verified with the original insurance application data information to verify whether the contents are consistent, the method further includes:

[0028] If the contents are inconsistent, re-extract the data string to be verified from the anti-counterfeiting QR code graphic of the insurance application form to be verified;

[0029] Check the structure of the data string to be verified to verify whether the data fields are complete and whether there are any missing fields;

[0030] If there are missing or incomplete fields in the data string to be verified, extract redundant data from the anti-counterfeiting two-dimensional code image, the redundant data including a hash value, and correct the data string to be verified based on the redundant data;

[0031] Compare the corrected data string to be verified with the original insurance application data information to verify again whether the contents are consistent;

[0032] If the contents are consistent, a pre-trained deep learning model is used to identify whether the target seal exists in the insurance application to be verified;

[0033] If the contents are inconsistent, the verification is confirmed to have failed, and a failure record is generated and sent to the administrator.

[0034] In one embodiment, converting the encrypted data string into two-dimensional code graphic data includes:

[0035] Encoding the encrypted data string, converting the encrypted data string into encoded data in a two-dimensional code format according to a two-dimensional code encoding standard;

[0036] Using a QR code generation tool to generate two-dimensional code graphic data according to the encoding data, the two-dimensional code graphic is composed of a two-dimensional matrix containing the encrypted data string;

[0037] The two-dimensional code graphic data is output as an image format for embedding into the insurance application form.

[0038] In one embodiment, the two-dimensional code graphic data is embedded into an insurance application form to generate an insurance application form with an anti-counterfeiting two-dimensional code graphic, including:

[0039] Confirm the type of the insurance application form, including electronic insurance application form and paper insurance application form;

[0040] When the insurance application form is an electronic insurance application form, obtaining the resolution of the display device, and determining the size of the QR code graphic based on the resolution;

[0041] Identify a blank area in the electronic insurance application page, and select a target blank area whose boundary exceeds a first preset value from the edge of the page and whose area meets the size requirement of the QR code image;

[0042] Embedding the two-dimensional code graphic into an independent display layer at the position of the target blank area to generate an electronic insurance application form with an anti-counterfeiting two-dimensional code graphic;

[0043] When the insurance application form is a paper insurance application form, the printing accuracy DPI of the printing device is obtained, and the size of the two-dimensional code graphic is determined based on the printing accuracy DPI;

[0044] Identify a blank area in the paper insurance application page, and select a target blank area whose border is more than a second preset value from the edge of the page and whose area meets the size requirement of the QR code image;

[0045] Embed the QR code image into the target blank area to generate a paper insurance policy with an anti-counterfeiting QR code graphic.

[0046] In one embodiment, the signature string is used as an encryption key, and the insurance application data information is used as the encrypted original text to perform an encryption operation to generate an encrypted data string, including:

[0047] Get the timestamp of the current encryption operation;

[0048] Combining the insurance application data information with the timestamp as the encrypted original text;

[0049] The signature string is used as an encryption key, and an encryption technology with a 256-bit key length is adopted to perform an encryption operation on the encrypted original text to generate an encrypted data string.

[0050] Furthermore, to achieve the above-mentioned purpose, the present invention also provides a QR code-based insurance form anti-counterfeiting device, the QR code-based insurance form anti-counterfeiting device comprising a memory, a processor, and a QR code-based insurance form anti-counterfeiting program stored in the memory and executable on the processor, the QR code-based insurance form anti-counterfeiting program, when executed by the processor, implements the steps of the QR code-based insurance form anti-counterfeiting method as described above.

[0051] Furthermore, to achieve the above-mentioned purpose, the present invention also provides a computer storage medium, on which a QR code-based insurance form anti-counterfeiting program is stored, and when the QR code-based insurance form anti-counterfeiting program is executed by a processor, the steps of the QR code-based insurance form anti-counterfeiting method as described above are implemented.

[0052] Beneficial effects: The present invention relates to an anti-counterfeiting method for an insurance application based on a QR code, which obtains the insurance application data information, including the insurance application number, product code, name of the insured, and the identification number of the insured; obtains the unique identity identifier of the business operator, and performs MD5 hash operation on the identifier to generate a signature string; uses the generated signature string as an encryption key, takes the insurance application data information as the encrypted original, performs encryption operation to generate an encrypted data string; converts the generated encrypted data string into QR code graphic data, and embeds the QR code graphic into the insurance application to form an insurance application with an anti-counterfeiting QR code graphic. The present invention effectively prevents the tampering and forgery of the insurance application content by encrypting the insurance application information and generating a QR code, ensures the security and authenticity of the data, improves the anti-counterfeiting performance of the insurance policy information, and reduces the workload of manual review, improves the review efficiency, and reduces the risks in the business process. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] The present invention will be further described below with reference to the accompanying drawings and embodiments, in which:

[0054] Figure 1 It is a flow chart of an embodiment of an anti-counterfeiting method for an insurance application form based on a QR code of the present invention;

[0055] Figure 2 It is a functional module schematic diagram of a preferred embodiment of the insurance application form anti-counterfeiting device based on a QR code of the present invention;

[0056] Figure 3 It is a structural schematic diagram of the device hardware operating environment involved in the implementation scheme of the insurance policy anti-counterfeiting device based on QR code of the present invention. DETAILED DESCRIPTION

[0057] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0058] In group property insurance business, the development of electronic processes is relatively lagging, and paper documents still account for a high proportion of the business process. The diversity and non-standardized formats of paper documents make it difficult to implement automated audit technology. The current audit process mainly relies on manual operations, which not only has a large workload and low efficiency, but also has limited audit coverage, making it difficult to ensure the authenticity of all insurance policies, increasing the risk of missed inspections and misjudgments in operations. This situation not only affects the efficiency of business processing, but also brings greater business compliance risks.

[0059] Although the existing OCR (optical character recognition) technology has certain applications in certain standardized scenarios, the actual application effect of OCR is limited in group property insurance due to the wide variety of products and ever-changing policy formats. Existing technology is difficult to conduct effective image recognition training for thousands of product categories, which greatly reduces the production value of this technology in policy review and cannot cover non-standardized documents such as homemade policies. These technical limitations make it difficult to automate policy review, and the current situation of relying on manual review cannot be effectively improved.

[0060] Furthermore, insufficient verification of the authenticity of the underwriting information, especially the insurance application, in group property insurance directly leads to potential exposure to claims risk. Once an insurance application with false information enters the underwriting process, the insurance company will face serious legal and economic risks in the claims process. If the insurance company cannot provide effective underwriting basis due to the authenticity of the insurance application, it may face a situation of losing the case, resulting in huge economic losses. In addition, the limitations of manual review will also delay the speed of claims decision-making, affect customer experience, and further increase the complexity and uncertainty of business processing.

[0061] With the development trend of digitalization and intelligence in the insurance industry, how to solve the problems of low efficiency of document review and insufficient authenticity verification in group property insurance business has become the key to breakthroughs in the industry.

[0062] See also Figure 1 , Figure 1 This is a flowchart of an embodiment of the QR code-based insurance application anti-counterfeiting method provided by the present invention. It should be noted that although the logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than that here.

[0063] like Figure 1 As shown, the insurance application form anti-counterfeiting method based on the QR code proposed by the present invention includes the following steps:

[0064] S10, obtaining insurance application data information, wherein the insurance application data information includes the insurance application number, product code, policyholder name and policyholder identification number;

[0065] In this embodiment, the obtained insurance application data specifically includes the following items:

[0066] Policy Number: This is a unique identifier for an insurance policy, usually automatically generated by the insurance system based on internal rules or sequences, and is used to track and manage a specific policy.

[0067] Product code: used to distinguish different insurance products. Each product has a corresponding code to ensure that the system can associate specific insurance products and related terms.

[0068] Name of the insured: This is the basic identity information of the insured. The system extracts it from the information filled in by the insured to ensure that it can match the insured's identity information in subsequent processes.

[0069] Policyholder identification number: This is the policyholder's unique identification number and is used for identity verification. The system verifies the legitimacy and validity of this information by connecting to an identity database or a third-party authentication system.

[0070] In the internal system of an insurance company, when a user initiates an insurance request, the system automatically generates a unique insurance policy number, which is usually based on a timestamp, a system internal counter, or a business serial number. This number is used to identify a specific insurance policy and is referenced in all subsequent related operations.

[0071] Product codes are pre-defined in the insurance product database, and different insurance products have unique codes. During the insurance application process, the system extracts the product code from the product database and binds it to the insurance application form.

[0072] The system extracts the policyholder's name and ID number from the personal information form submitted by the policyholder. The system will connect to an identity authentication service (such as a national identity database or a third-party identity authentication service) to ensure that the information provided by the policyholder is accurate and prevent false identities.

[0073] In other specific embodiments, the system combines artificial intelligence technology to optimize the extraction and verification process of insurance application data information. When obtaining insurance application data information, the system automatically parses and extracts key information in unstructured forms (such as insurance application number, product code, insured name and identification number in PDF or scanned images) through natural language processing (NLP) technology. Through machine learning algorithms, the system automatically trains models to identify insurance applications in different formats and improves the accuracy of data extraction. This method enables the system to automatically process complex and diverse insurance application formats, effectively improving processing efficiency and accuracy.

[0074] The system automatically obtains the insurance application number, product code, name of the insured and identity number, which can ensure the accuracy and completeness of the insurance application data information, avoid manual input errors, and reduce the risk of information tampering. It improves the efficiency and security of data processing, provides a reliable data foundation for subsequent data encryption, verification and anti-counterfeiting, thereby enhancing the anti-counterfeiting and credibility of the insurance application in the entire insurance business process, and reducing the business risks caused by false information.

[0075] S20, obtaining a unique identity identifier of the service operator, performing an MD5 hash operation on the unique identity identifier, and generating a signature string;

[0076] In this embodiment, the unique identity identifier of the business operator is the unique identity information used by the system to identify the business operator who performs a specific operation, which is usually the operator's user name, work number or other identifier. This identity identifier can be obtained in the following ways:

[0077] Extract after logging into the system: The user name or work number entered by the operator when logging into the system, the system will automatically record the information and store it in the background for subsequent operation authentication and encryption.

[0078] Assigned by the system: After the operator enters the business system, the system dynamically generates a unique identity identifier for him / her to ensure that each operation can be traced back to the operator.

[0079] The system extracts the operator's identity identifier when he logs in and processes it according to the pre-set identifier generation rules. The identifier is obtained through database query, authentication management system or session ID. The system performs MD5 operation on the obtained identity identifier by calling the hash algorithm function. The MD5 algorithm converts the input data into a 128-bit hash value to ensure that the generated signature string is of fixed length and cannot be directly restored to the original data. After completing the MD5 operation, the signature string generated by the system will be used for subsequent encryption operations to ensure that the association between the operator's identity and data is safe and reliable.

[0080] In one specific implementation, the system obtains the unique identifier of the business operator not only from the user name or work number, but also from the operator's login device ID and geographic location information. These additional identifiers serve as an additional layer of identity authentication, and all identity identifiers are combined into a string for encryption before the MD5 hash operation. This method enhances the reliability of operator identity authentication and prevents improper use.

[0081] By performing MD5 hashing on the unique identity identifier of the business operator to generate a signature string, the operator's identity information is effectively protected during data encryption and transmission. The MD5 hashing operation makes the identity information irreversible, enhances the security of the operator's identity and the system's anti-tampering capabilities, and provides a reliable encryption key for subsequent encryption operations.

[0082] S30, using the signature string as an encryption key and the insurance application data information as the encrypted original text, performing an encryption operation to generate an encrypted data string;

[0083] In this embodiment, the generated signature string has been converted into a 128-bit fixed-length string through the MD5 hash algorithm. The signature string is used as the core key for encryption operations to encrypt the insurance policy data information. The main function of the signature string is to provide a unique and secure encryption key to ensure that the encryption process of each operator is independent, avoid using the same key for multiple encryptions, and reduce the risk of information leakage. After the system generates the MD5 signature string, it is provided as an input parameter to the encryption module. The generation and storage of the signature string is performed in the memory to ensure that it will not be leaked externally before the encryption operation is completed. In the encryption operation, the system uses the signature string as a key and passes it to the key parameter part in the encryption algorithm (such as AES, DES, etc.) to encrypt the original data.

[0084] The insurance policy data information, including the insurance policy number, product code, policyholder name and policyholder identification number, is the original data for encryption. The system needs to ensure that this data is not tampered with or leaked during transmission or storage, so it is used as the encrypted original text for encryption operations. The combination and arrangement order of the insurance policy data information are based on business needs, usually through splicing or serialization technology, to package multiple data fields into an overall string or data block. The system combines the extracted insurance policy data into an encrypted original text through string splicing, serialization or JSON formatting. The definition of the data structure depends on the business needs in the insurance policy system. For example, the insurance policy number and product code may need to be arranged in order with the name and ID card information, and a specific delimiter is used to connect these data to ensure that they can be correctly restored during the encryption and decryption process.

[0085] The core of the encryption operation is to combine the aforementioned encryption key (signature string) and the encrypted original text (insurance policy data information) and encrypt them using a symmetric encryption algorithm (such as AES256). The encrypted data string is ciphertext and cannot be directly read or restored to the original data, ensuring the security of the insurance policy data. The generated encrypted data string will be used for subsequent anti-counterfeiting and verification operations. The system calls the encryption algorithm library (such as AES or DES) and uses the encrypted original text and signature string as input parameters for encryption processing. After the encryption algorithm is executed, the system outputs an encrypted data string. The data string is a ciphertext of fixed length, which is irreversible and can only be decrypted with the same key. The encryption process is performed in the system's security module to prevent external leakage of encryption keys and data.

[0086] In another specific implementation, the system uses segmented encryption to encrypt different fields in the insurance application data information separately. For example, the insurance application number and product code are used as the first segment of data, and the signature string is used for AES encryption; the insured's name and identity number are used as the second segment of data, and the random key generated by the system is used for encryption. Finally, the encrypted segments of data are merged into an encrypted data string. Segmented encryption can enhance the flexibility of data protection and further improve data security.

[0087] By using the signature string as the encryption key and the insurance policy data information as the encrypted original text for encryption operations, the core data of the insurance policy can be effectively protected during transmission and storage. The encrypted data string is irreversible, preventing information from being tampered with and leaked, significantly improving the anti-counterfeiting and confidentiality of the insurance policy data, and providing a solid foundation for subsequent anti-counterfeiting verification and information processing.

[0088] S40, converting the encrypted data string into two-dimensional code graphic data, embedding the two-dimensional code graphic data into the insurance application form, and generating an insurance application form with an anti-counterfeiting two-dimensional code graphic.

[0089] In the present embodiment, the encrypted data string is converted into a QR code format through a specific QR code generation tool or API. A QR code is a two-dimensional graphic that can represent a string of encrypted data in a visual form. This operation needs to encode the data according to the encoding standard of the QR code (such as the QR code standard) to ensure that the QR code can be correctly decoded and restored to the original encrypted data string when scanned. The system calls a QR code generation tool or library function (such as Google Zxing or other open source tools) to input the encrypted data string into the generation function. The generation tool encodes the encrypted data into a QR code according to a predefined fault tolerance level (such as L, M, Q, H). In the process of QR code generation, the size of the graphic, the amount of data storage, and the fault tolerance level can also be selected according to business needs to adapt to different application scenarios.

[0090] The generated QR code graphic data is a two-dimensional graphic containing encrypted data, which needs to be embedded in the appropriate location of the insurance application form, usually in the page of the electronic insurance application form or the paper insurance application form. The embedding operation needs to ensure that the QR code does not visually affect the other contents of the insurance application form while maintaining its scannability. The system embeds the QR code graphic file into the PDF file or paper document template of the insurance application form through a document generation tool (such as a PDF generator or a document processing tool). The embedded location can be the upper right corner, lower right corner or other blank area of ​​the insurance application form. During the embedding process, the system will adjust the size and resolution of the QR code to ensure that it can be read normally on different devices.

[0091] The generated insurance application not only contains the standard insurance information, but also contains an encrypted data string embedded in the form of a QR code. The QR code has an anti-counterfeiting function. By scanning the QR code, the encrypted data string can be decoded to verify the authenticity of the insurance application. After the QR code is embedded, the insurance application will be generated as a PDF file or ready to print. For electronic insurance applications, the file will be stored in the system for downloading or sending. For paper insurance applications, the system is ready for printing and ensures that the QR code is clear enough for scanning after printing.

[0092] In another specific implementation, after the QR code graphic is generated, the system further optimizes the display effect of the QR code through image processing technology. First, the generated QR code graphic is combined with the brand logo of the insurance application form to form a unique anti-counterfeiting QR code with the corporate logo. The QR code is embedded in the blank area of ​​the paper insurance application form. Before printing, the system adjusts the resolution of the QR code to meet the resolution requirements of the printer, ensuring that the printed QR code can maintain high-quality readability.

[0093] In other specific implementations, the system uses dynamic QR code technology. Each generated QR code not only contains an encrypted data string, but also comes with a timestamp or unique identification code to ensure that each generated QR code is unique. After the QR code is embedded in the insurance application form, the system can verify its generation time and usage status by scanning the QR code to prevent the QR code from being copied or tampered with. This implementation is mainly used in high-security scenarios to ensure the anti-counterfeiting of the insurance application form.

[0094] By converting the encrypted data string into a QR code and embedding it into the insurance application, the forgery and tampering of the insurance application information can be effectively prevented. As an anti-counterfeiting measure, the QR code can quickly verify the authenticity of the insurance application by scanning, while ensuring the security of data during transmission and storage. It not only improves the security of the insurance application information, but also simplifies the verification process and improves the efficiency of business processing.

[0095] The present invention relates to an anti-counterfeiting method for an insurance application form based on a two-dimensional code, which comprises obtaining insurance application form data information, including an insurance application form number, a product code, an insured person's name and an insured person's identification number; obtaining a unique identity identifier of a business operator, and performing an MD5 hash operation on the identifier to generate a signature string; using the generated signature string as an encryption key, taking the insurance application form data information as an encrypted original, performing an encryption operation to generate an encrypted data string; converting the generated encrypted data string into two-dimensional code graphic data, and embedding the two-dimensional code graphic into the insurance application form to form an insurance application form with an anti-counterfeiting two-dimensional code graphic. The present invention effectively prevents tampering and forgery of the insurance application form content by encrypting the insurance application form information and generating a two-dimensional code, ensures the security and authenticity of the data, improves the anti-counterfeiting performance of the insurance application information, and reduces the workload of manual review, improves the review efficiency, and reduces the risk in the business process.

[0096] In one embodiment, in the above S40, after the insurance application form with the anti-counterfeiting two-dimensional code graphic is generated, the following further includes:

[0097] S401, obtaining an insurance application form to be verified, and extracting a data string to be verified from the anti-counterfeiting two-dimensional code graphic of the insurance application form to be verified;

[0098] S402, decrypting the data string to be verified to restore the data information to be verified;

[0099] S403, identifying the insurance application number from the data information to be verified, and confirming the corresponding original insurance application data information based on the insurance application number;

[0100] S404, comparing the data information to be verified with the data information of the original insurance application to verify whether the contents are consistent;

[0101] S405, if the contents are consistent, identifying whether the target seal exists in the insurance application to be verified through a pre-trained deep learning model;

[0102] S406: If the target seal is identified, confirm that the insurance application to be verified has passed the verification.

[0103] In this embodiment, the anti-counterfeiting QR code graphic on the insurance application to be verified is read through a QR code scanning device or application. The data in the QR code contains encrypted information, and the system extracts the data as a data string to be verified for subsequent decryption and verification operations. The system reads the QR code graphic data on the insurance application through a QR code recognition device such as a mobile phone application or a scanner. The QR code recognition tool parses the data in the QR code into an encrypted data string. The data string will be passed to the decryption module in the system as input for subsequent verification steps.

[0104] The data string to be verified has been encrypted, so the original insurance policy data information needs to be restored through a decryption algorithm. The decryption process uses the same key and algorithm as the encryption to ensure that the data can be accurately restored. The system calls the decryption module and uses the original signature string or key to execute the decryption algorithm corresponding to the encryption (such as AES decryption). Through this process, the system can restore the encrypted data string to the plaintext insurance policy information (including insurance policy number, product code, policyholder name, ID number, etc.).

[0105] The decrypted data contains the insurance application number, which the system uses to search and match the original insurance application data information in the insurance application database. This ensures that the decrypted data string corresponds to the original insurance application stored in the database. The system extracts the insurance application number from the decrypted data and searches the system database based on the number to find the corresponding original insurance application data. The query results will be extracted and used to compare with the decrypted data to ensure the accuracy of the data.

[0106] The system compares the decrypted insurance application data with the original insurance application data in the database to check whether the two are consistent and verify the authenticity of the data. If the data is consistent, it means that the insurance application has not been tampered with. The system checks the decrypted data (such as insurance application number, product code, insured person's name, ID number) item by item through field comparison to see whether it is consistent with the original insurance application data. After the comparison is completed, the system generates a verification result.

[0107] The system uses a deep learning model to identify the seal in the insurance application image. The model is pre-trained with a large number of seal samples and can identify specific types of seals and ensure that the seal used on the insurance application is a legitimate seal. The system extracts the seal area in the insurance application image and inputs it into the pre-trained deep learning model. The model will identify the seal based on its shape, texture and other features to determine whether the seal is the target seal. The system uses the output of the model as a basis for verification.

[0108] If the system recognizes the target seal and the data is consistent, the system confirms that the insurance policy is authentic and legal, and completes the verification process. After completing the comparison and seal recognition, the system automatically generates a verification result and records the detailed information of the verification, confirming that the insurance policy has passed the verification.

[0109] In other specific implementations, the system uses an advanced security verification mechanism to combine QR code data and external data sources for multi-dimensional comparison. After extracting the data string to be verified, the system not only decrypts the insurance application data, but also accesses an external authentication service to perform a secondary verification of the policyholder information. After the verification is completed, the anti-counterfeiting seal in the insurance application is identified through a deep learning model to confirm the authenticity of the insurance application.

[0110] This embodiment extracts the encrypted data string in the QR code and decrypts it, and combines the original data comparison of the database and the seal recognition technology to efficiently verify the authenticity of the insurance application form and prevent the insurance application form from being tampered with or forged. The use of deep learning models to automatically recognize seals further improves the accuracy of verification, simplifies the manual review process, and reduces the risk of errors in operations.

[0111] In one embodiment, in the above S405, identifying whether the target seal exists in the insurance application to be verified by using a pre-trained deep learning model includes:

[0112] S4051, extracting a seal image from the insurance application form to be verified, performing noise removal and binarization processing on the seal image to separate the seal area from the background;

[0113] S4052, inputting the processed seal image into a pre-trained deep learning model to identify the seal type in the seal image, where the seal type includes a company seal and a business-specific seal;

[0114] S4053, comparing the identified seal type with the type of the target seal;

[0115] S4054, if the seal type does not match, confirm that the target seal does not exist in the insurance application to be verified;

[0116] S4055, if the seal type matches, extracting feature information corresponding to the identified seal type from the seal image, the feature information including the shape and texture of the seal;

[0117] S4056, matching the characteristic information with the characteristics of the target seal, and analyzing the similarity between the seal image and the target seal;

[0118] S4057: If the similarity exceeds the similarity threshold, confirm that the target seal exists in the insurance application to be verified.

[0119] In this embodiment, the seal area in the insurance application form is extracted from the entire image through image processing technology. Noise removal is used to remove interference in the seal image, such as background texture, shadows, etc. Binarization converts the image into black and white mode to ensure clear separation of the seal and the background, making the subsequent recognition process more accurate. The system first uses an image processing algorithm (such as median filtering) to remove noise in the insurance application form image. Then, the image is binarized by an adaptive threshold method to convert the image into a black and white pixel image to ensure that the seal part is clearly distinguished from the background. At this point, the seal area is successfully separated and ready for further processing.

[0120] By inputting the processed image into a pre-trained deep learning model, the system can automatically identify the type of seal. The deep learning model is trained based on a large number of seal samples and can identify different types of seals, such as company seals and business-specific seals. The system inputs the seal image after noise removal and binarization into the deep learning model. The model extracts the features of the image through a convolutional neural network (CNN) and identifies the type of seal based on previously trained samples. The output result is the type of seal, such as company seal or business-specific seal.

[0121] The system compares the identified seal type with the expected target seal type. It is mainly used to verify whether the seal meets expectations, such as confirming that a certain insurance application should be a company seal, but the identified seal is another type. The system compares the identified seal type with the target seal type through business logic or database preset rules. If the two are consistent, it means that the insurance application uses the correct seal. If they are not consistent, it means that there may be problems with the insurance application.

[0122] If the identified seal type does not match the target seal type, the system confirms that the expected seal is not used on the insurance application and marks the verification as failed. After comparing the seal type, if the system finds a mismatch, it automatically marks the insurance application as failed, outputs an error message, and sends the verification result to the management system or operator for further processing.

[0123] In the case of a seal type match, the system further extracts the shape and texture features of the seal from the image to confirm the authenticity of the seal. This process further verifies the legitimacy of the seal by analyzing the seal's outline, detailed texture and other feature information. The system obtains the seal's shape, texture and other feature information through the feature extraction layer of the convolutional neural network. Features include the seal's outer contour, the arrangement of the internal text and patterns, etc. The system stores this feature information and uses it for subsequent similarity comparisons.

[0124] The system compares the extracted seal features with the preset features of the target seal, and determines whether the seal is the target seal by calculating the similarity between the two. The system compares the extracted seal image features with the target seal features in the database. The similarity of the seals is calculated using an image feature matching algorithm (such as the structural similarity index, SSIM). If the similarity exceeds a preset threshold, the system considers the seal to match the target seal.

[0125] The system confirms that the seal in the insurance application is the target seal based on the similarity result, and then completes the verification of the insurance application. The system sets a similarity threshold (such as 90%). When the similarity between the seal image and the target seal exceeds the threshold, the system confirms that the insurance application has been verified and outputs the result of successful verification. The system also records the verification result for subsequent review.

[0126] In one specific implementation, the entire image of the insurance policy to be verified is scanned to find the area where the seal is located, and the seal area is automatically identified by a region growing algorithm. The region growing algorithm is a technique for image segmentation that starts from an initial seed point and expands the region by the similarity of adjacent pixels until the entire seal area is segmented. The specific steps are as follows:

[0127] The initial seed point of the stamp area can be selected based on color, texture or shape features. Check the properties of adjacent pixels (such as color value or grayscale value), and if they are similar to the properties of the seed point, add the pixel to the region. Continue to expand the region until all similar pixels are included to form a complete stamp area. After the region growing is completed, the system uses the extracted stamp area as input for the next step of processing.

[0128] After extracting the seal image, the system uses feature extraction technology based on local binary patterns (LBP) to remove noise and processes the binary image in combination with a region filling algorithm to ensure the integrity of the seal.

[0129] LBP is a commonly used image processing method for extracting texture features of images. By comparing the grayscale value of each pixel with its neighboring pixels, a binary pattern is generated to form a local texture descriptor. The specific steps are as follows:

[0130] For each pixel, compare it with the grayscale value of the surrounding pixels. If the grayscale value of the adjacent pixel is greater than or equal to the central pixel, assign a value of 1; otherwise, assign a value of 0. Based on these comparison results, a binary pattern is generated to represent the local texture of the pixel. The system encodes the LBP features of the entire image to extract the texture information of the image, while eliminating local noise and maintaining the integrity of the stamp area.

[0131] In the process of LBP feature extraction, the local binary pattern can naturally filter out some high-frequency noise information, because the grayscale fluctuation of noise points in a small range cannot significantly affect the binarization result. Ultimately, the image processed by LBP will have clearer seal contours and texture features.

[0132] After noise removal, some small holes or broken areas may appear. To ensure the integrity of the seal, the system uses a region filling algorithm to process the binary image. Region filling ensures the continuity of the seal by identifying isolated pixel areas in the image and connecting them with adjacent areas. The filling process usually includes the following steps:

[0133] Find the holes or broken areas in the image. Fill the holes according to the grayscale or binary information of the adjacent pixels to keep the shape of the seal intact. After the area is filled, the system uses a fixed or adaptive threshold method to binarize the image and convert the seal image into a black and white image. After binarization, the seal is completely separated from the background area, providing a basis for subsequent feature extraction and analysis.

[0134] The processed seal image is used as input to the pre-trained deep learning model. The model is trained with a large amount of seal sample data and can distinguish different types of seals, such as company seals, business-specific seals, etc. In order to adapt to dynamic environments and different image qualities, the system uses reinforcement learning (RL) to optimize the parameters of the deep learning model. Reinforcement learning optimizes the model through the following steps:

[0135] The system sets a goal, such as improving the accuracy of seal recognition. Each time recognition is performed, the system obtains the performance of the model through environmental feedback and adjusts the model parameters based on the feedback. Through repeated adjustments, the model can adaptively adjust its parameters when processing different images, thereby improving the accuracy and robustness of seal recognition.

[0136] After comparing the seal types, the system uses a texture feature analysis method based on the gray-level co-occurrence matrix (GLCM) to conduct an in-depth analysis of the seal texture, and finally determines the similarity of the seals through frequency domain analysis technology to judge their authenticity.

[0137] Gray-level co-occurrence matrix (GLCM) is a commonly used texture analysis method that can extract the relationship between different gray levels from an image. By calculating the number of co-occurrences of gray values ​​in specific directions (such as 0°, 45°, 90°), a co-occurrence matrix is ​​generated, and texture features such as contrast, correlation, entropy, etc. are extracted from it. The system generates a corresponding co-occurrence matrix based on the gray value of the seal image. Multidimensional texture features are extracted from the co-occurrence matrix to characterize the overall texture structure of the seal. By analyzing the texture features of GLCM, the system can deeply interpret the texture information of the seal. This step can capture subtle texture differences in the seal, which is used to verify the authenticity and uniqueness of the seal.

[0138] The system converts the image from the time domain to the frequency domain through Fourier transform and analyzes the frequency characteristics of the seal. Frequency domain analysis can capture the global structure and high-frequency details in the seal image, and is particularly suitable for processing images with complex textures. Through Fourier transform, the grayscale value of the seal image is converted into a frequency distribution. Frequency characteristics, such as the high-frequency and low-frequency components of the image, are identified and compared with the frequency characteristics of the target seal.

[0139] The system compares the texture features extracted by GLCM and the frequency features obtained by frequency domain analysis with the pre-stored features of the target seal. The similarity between the seal to be verified and the target seal is calculated using methods such as structural similarity index (SSIM) or cosine similarity. If the similarity exceeds the preset threshold, the system confirms that the seal is the target seal.

[0140] In this embodiment, by performing noise removal, binarization processing and seal recognition on the seal image using a deep learning model, the system can accurately identify the type of seal on the insurance policy and further verify the authenticity of the seal through similarity analysis of features such as shape and texture.

[0141] In one embodiment, after the above S404, the method further includes:

[0142] S4041, if the contents are inconsistent, re-extracting the data string to be verified from the anti-counterfeiting two-dimensional code graphic of the insurance application form to be verified;

[0143] S4042, checking the structure of the data string to be verified, verifying whether the data fields are complete, and checking whether there are any missing fields;

[0144] S4043, if there are missing fields or incomplete fields in the data string to be verified, extract redundant data from the anti-counterfeiting two-dimensional code image, the redundant data including a hash value, and correct the data string to be verified based on the redundant data;

[0145] S4044, comparing the corrected data string to be verified with the original insurance application data information to verify again whether the contents are consistent;

[0146] S4045, if the contents are consistent, identifying whether the target seal exists in the insurance application to be verified through a pre-trained deep learning model;

[0147] S4046: If the contents are inconsistent, the verification is confirmed to have failed, and a failure record is generated and sent to the administrator.

[0148] In this embodiment, if the decrypted data content is inconsistent with the original insurance application data, the encrypted data string is re-extracted from the QR code on the insurance application to be verified. The system scans the QR code graphic on the insurance application and uses the QR code decoding tool to extract the encrypted data string. The data string will be used in the subsequent structural inspection and correction steps.

[0149] The extracted data string is checked for integrity to ensure that all necessary fields exist and the data format meets the requirements. If the data structure is incomplete or fields are missing, subsequent verification may be inaccurate. The system checks the fields in the data string one by one through a preset template or data structure model. Each field (such as insurance policy number, product code, etc.) is compared according to predefined rules to verify whether the field is complete and meets the data format requirements. If missing or incomplete fields are found, the system will mark these issues and enter the next step of the data correction process.

[0150] In the case of missing or incomplete data fields, the system extracts redundant data from the QR code image. Redundant data usually includes auxiliary information such as hash values, which can help recover or correct missing fields. The hash value is a unique summary of the original data. By comparing the hash value, the system can find and correct errors in the data string. When the system extracts the QR code, it will also parse the redundant data in the QR code. Redundant data is usually stored in a special area of ​​the QR code or implemented through multi-level encoding. The system verifies the accuracy of the current data string based on the hash value of the redundant data, and finds and corrects the fields with errors or missing. The correction process uses redundant information to restore the data to make it consistent with the original data.

[0151] Based on the mismatch of the hash values, the system infers possible missing fields (such as the insurance policy number) and decides to try to correct the field through exhaustive or known data sets. The system first generates multiple candidate numbers according to the preset numbering rules or possible values ​​in the historical records. For each candidate number, the system combines it with other extracted complete data to form a new complete data string and hashes the data string. The system compares the newly generated hash value with the redundant hash value in the QR code.

[0152] When the hash value generated by a candidate number matches the redundant hash value in the QR code, the system confirms that the candidate number is the correct insurance policy number. At this time, the system fills the number into the data to be verified and completes the data correction. The corrected data is hashed again, and the system confirms that the hash value completely matches the redundant hash value in the QR code. The verification is successful, indicating that the data has been restored and is complete.

[0153] The corrected data string needs to be compared with the original insurance application data again to ensure the integrity of the corrected data. If the comparison results are consistent, it means that the data has been successfully corrected. The system uses the same comparison mechanism to compare the corrected data with the original insurance application data in the database field by field. The comparison process includes consistency verification of data content, format, and structure to ensure that the data is accurately repaired and its legitimacy is restored.

[0154] If the corrected data is successfully matched, the system will continue to identify the seal in the insurance application through the pre-trained deep learning model to verify the legitimacy of the insurance application. The model is trained based on a large number of seal samples and can identify the shape and texture features of the target seal. The system extracts the seal image on the insurance application and inputs it into the pre-trained deep learning model, which analyzes the shape and texture of the seal through a convolutional neural network (CNN). The model outputs the seal type and similarity, and the system determines whether the seal is the target seal based on the similarity.

[0155] If the verification still fails after correcting the data, the system will confirm that the policy verification has failed and generate a related failure record. This record will be automatically sent to the management staff to remind them to conduct further manual verification or take corrective measures. The system will automatically generate a verification failure report, including detailed information such as the reason for failure, data comparison results, and error fields. The generated record will be sent to the management system and the relevant operators or managers will be notified to take follow-up actions.

[0156] This embodiment can effectively deal with data missing or inconsistent situations by re-extracting the data string from the QR code and performing data integrity check and correction. The introduction of redundant data improves the accuracy of data recovery, and the application of deep learning models in seal verification further enhances the automation and accuracy of insurance application verification. It reduces the workload of manually processing erroneous data and reduces the risk of errors in the verification process.

[0157] In one embodiment, in the above S40, converting the encrypted data string into two-dimensional code graphic data includes:

[0158] S407, encoding the encrypted data string, converting the encrypted data string into encoded data in a two-dimensional code format according to a two-dimensional code encoding standard;

[0159] S408, using a QR code generation tool to generate two-dimensional code graphic data according to the encoded data, wherein the two-dimensional code graphic is composed of a two-dimensional matrix including the encrypted data string;

[0160] S409, outputting the two-dimensional code graphic data into an image format for embedding into the insurance application form.

[0161] In this embodiment, the encrypted data string is encoded to ensure that it complies with the encoding standard of the two-dimensional code (such as the QR code standard). The two-dimensional code encoding standard specifies how to convert data into binary information that can be represented as a two-dimensional code graphic. This encoding process ensures that the data can be correctly converted and represented in the two-dimensional code. The system calls the encoding tool to convert the encrypted data string into an encoding format acceptable to the two-dimensional code. This encoding process structures the data according to the two-dimensional code standard, usually including data compression and error correction (such as the four fault tolerance levels of L, M, Q, and H in the QR code) to ensure that the two-dimensional code can be successfully read even if it is partially damaged.

[0162] Use the QR code generation tool to generate a QR code graphic from the encoded data string. QR code is essentially a two-dimensional matrix composed of a group of black and white modules, each module represents the binary form of the encoded data. The size, fault tolerance level, data capacity and other parameters of the QR code can be set through the generation tool. The system uses built-in or external QR code generation tools (such as Google Zxing, QRCode.js, etc.), inputs the encoded data into the tool, and generates the corresponding QR code graphic. The QR code generation tool will automatically adjust the size and structure of the QR code according to the set fault tolerance level and data volume. In the end, the generated QR code graphic data can contain several thousand bytes of encrypted data strings and be displayed in the form of a matrix.

[0163] The generated QR code graphic needs to be output to a standard image format (such as PNG, JPG, SVG, etc.) so that it can be subsequently embedded in an electronic or paper insurance application form. This process ensures that the QR code can be stored and used as an image file. The system saves the generated QR code graphic data in a common image file format. The output image file can be embedded in a PDF file or web page for display or printing. The format and resolution of the image file can be adjusted according to different application scenarios to ensure that the QR code can be correctly scanned and decoded.

[0164] In another specific implementation, in order to enhance the security and traceability of the QR code, the system uses multi-layer encryption technology. Before encoding the encrypted data string into a QR code, the system uses double-layer encryption: the first layer uses a symmetric encryption algorithm (such as AES), and the second layer uses an asymmetric encryption algorithm (such as RSA). This double-layer encrypted data string is encoded in multiple layers to generate a QR code. The QR code generation tool selects an H-level fault tolerance level based on the complexity of the data to ensure that even if the QR code is damaged over a large area, the data can be restored through redundant information. The generated QR code graphics are output in SVG format. The vector graphics of SVG ensure that the QR code can be flexibly displayed on different devices or resolutions, and is suitable for embedding into electronic contracts or dynamic web pages.

[0165] By converting the encrypted data string into a QR code and embedding it into the insurance application, the system can effectively protect the security of the insurance application information and ensure that the data is not easily tampered with during transmission and display. The QR code generation process complies with standardized coding rules and has efficient fault tolerance. Even if the QR code is partially damaged, the readability of the data can still be guaranteed. At the same time, the QR code is output in an image format, which is easy to embed into an electronic or paper insurance application, realizes data anti-counterfeiting verification, and improves the efficiency and reliability of insurance application processing.

[0166] In one embodiment, in the above S40, the two-dimensional code graphic data is embedded into the insurance application form to generate the insurance application form with the anti-counterfeiting two-dimensional code graphic, including:

[0167] S410, confirming the type of the insurance application form, including an electronic insurance application form and a paper insurance application form;

[0168] S411, when the insurance application form is an electronic insurance application form, obtaining a resolution of a display device, and determining a size of the QR code graphic based on the resolution;

[0169] S412, identifying a blank area in the electronic insurance application page, and selecting a target blank area whose boundary exceeds a first preset value from the page edge and whose area meets the size requirement of the QR code image;

[0170] S413, embedding the two-dimensional code graphic into an independent display layer at the position of the target blank area to generate an electronic insurance application form with an anti-counterfeiting two-dimensional code graphic;

[0171] S414, when the insurance application form is a paper insurance application form, obtaining the printing precision DPI of the printing device, and determining the size of the two-dimensional code graphic based on the printing precision DPI;

[0172] S415, identifying a blank area in the paper insurance application page, and selecting a target blank area whose boundary is greater than a second preset value from the page edge and whose area meets the size requirement of the QR code image;

[0173] S416, embed the QR code image into the target blank area to generate a paper insurance application form with an anti-counterfeiting QR code graphic.

[0174] In this embodiment, the system first needs to confirm the type of insurance application form according to the business process, whether it is an electronic insurance application form or a paper insurance application form. Different types of insurance applications have different requirements in the QR code generation and embedding process. The system determines the type of insurance application form through business logic or user selection. If it is an electronic insurance application form, the file processed is a PDF or web page format; if it is a paper insurance application form, the system processes a static file format for printing.

[0175] When the insurance application form is an electronic one, the size of the QR code needs to be adjusted according to the screen resolution of the device used by the user. Different screen resolutions affect the display effect of the QR code. The size of the QR code must be compatible with the display device to ensure that the QR code is still clear and readable at different resolutions. The system obtains the resolution of the current device (such as 1920x1080, 2560x1440, etc.) by calling the device API. Based on the display resolution, the system will calculate the optimal display size of the QR code to ensure that it is neither blurry nor too large to affect the display of other page content.

[0176] The system needs to find a blank area in the electronic insurance application form that is suitable for placing the QR code, and the area of ​​this area must be large enough to ensure that the QR code can be embedded normally and remain clear. This step needs to ensure that the QR code does not block other important content on the page. The system scans the insurance application page through a page analysis tool, identifies blank areas in the page, and selects a suitable area based on preset standards (such as margin requirements and minimum QR code area). This area is usually located in the lower right or lower left corner of the page, and is more than a first preset value (for example, 20 pixels) away from the edge of the page to ensure the rationality of the overall layout of the page.

[0177] When a QR code is embedded in an insurance application form, if it is too close to the edge of the page, it may block or cover important information on the page (such as text, graphics, etc.). Setting margins can ensure that the QR code is kept at an appropriate distance from other content to avoid interfering with the page layout. At the same time, most QR code scanning devices or applications require a certain amount of blank area (also known as "quiet zone" or "static margin") around the QR code to correctly recognize the QR code. Margins can ensure that there is enough blank space around the QR code to improve the recognition efficiency of the scanning device. If the QR code is too close to the edge of the page, it may not be possible to read the QR code information completely. Especially in paper insurance applications, if there is not enough blank space around the QR code, the edges of the QR code may be blurred or damaged when printed. Setting margins can help avoid resolution loss due to limitations of the printing device and ensure that the QR code is clear and scannable.

[0178] After identifying the appropriate blank area, the QR code graphic needs to be embedded in an independent display layer of the electronic insurance application form. This operation ensures that the QR code can be independent of other page elements to prevent overlapping with other content, which would affect the readability of the QR code. The system creates an independent display layer in the PDF or web page file (such as an image layer in PDF or a div layer in HTML), and embeds the generated QR code graphic into the display layer to ensure that the position of the QR code is fixed and does not affect the display of other content on the page. The independent display layer separates the QR code from other page content (such as text, charts, etc.) to prevent the QR code graphic from overlapping or conflicting with the page content. This helps keep the page content clear and readable, and ensures that the QR code is always scannable. For example, in HTML, the QR code is placed in an independent The QR code will not be correctly recognized due to overlapping page elements, which will not affect other texts or pictures.

[0179] When the insurance application form is a paper version, the system needs to adjust the size of the QR code according to the resolution (DPI, dots per inch) of the printing device. Different printing devices have different DPIs, and the QR code must be adapted according to the printing accuracy to ensure that the printed QR code remains clear and scannable. The system obtains the DPI value of the current printing device by accessing the driver or API of the printing device. Depending on the DPI, the system will dynamically adjust the resolution and size of the QR code image. For example, the higher the DPI, the size of the QR code can be reduced accordingly while still remaining clear and readable.

[0180] Similar to the electronic insurance application form, the system needs to find a blank area in the paper insurance application form that is suitable for embedding the QR code. The area of ​​​​the area needs to be large enough and the distance from the edge of the page exceeds the preset value so that the printed QR code can be correctly scanned without affecting other content. The system will analyze the paper insurance application page to be printed and identify blank areas suitable for embedding the QR code. Usually, the system will select a relatively blank area at the bottom of the page or at the edge of the page to ensure that the QR code does not affect the printed text or other important information. The selection of the area needs to follow the preset distance (for example, more than 10mm from the edge) and area standards.

[0181] After determining the target blank area, the system will embed the generated QR code graphic into the printing template of the paper insurance application. The QR code image needs to be adapted according to the size of the blank area and embedded in the specified position in the template to ensure that the printed paper insurance application has the anti-counterfeiting QR code function. The system inserts the QR code image into the specified blank area of ​​the paper insurance application through the document editing tool or the printing template processing tool. At this point, the QR code has been resized according to the DPI and page layout of the printing device to ensure that the QR code can be clearly and accurately displayed on the insurance application after printing, and finally generate a paper insurance application with an anti-counterfeiting QR code graphic.

[0182] This embodiment dynamically adjusts the size of the QR code and embeds it in the appropriate page position according to different insurance application forms. The system can ensure the clarity and readability of the QR code while maintaining the integrity and beauty of the page content. Whether it is an electronic insurance application form or a paper insurance application form, the embedding of the QR code can effectively improve the anti-counterfeiting ability of the insurance application form and ensure data security and reliability.

[0183] In one embodiment, the above S30 includes:

[0184] S301, obtaining the timestamp of the current encryption operation;

[0185] S302, combining the insurance application data information with the timestamp as an encrypted original text;

[0186] S303, using the signature string as an encryption key, adopting encryption technology with a 256-bit key length, encrypting the encrypted original text to generate an encrypted data string.

[0187] In this embodiment, when performing encryption operations, the system will obtain the current timestamp (usually in seconds or milliseconds). The timestamp is used to increase the uniqueness of each encryption operation. Even if the same data is encrypted repeatedly, the generated encryption results will be different due to the change of the timestamp. As part of the dynamic input, the timestamp ensures that the encrypted data has time relevance to prevent the data from being replayed or tampered with.

[0188] The system obtains the current precise timestamp by calling the time API. The timestamp indicates the exact time point when the encryption operation occurs (such as the number of seconds since January 1, 1970). The timestamp can be used as part of the encrypted original text and encrypted together with the insurance policy information. Specific calls include currentTimeMillis() (Java) or Date.now() (JavaScript).

[0189] The data information of the insurance application is combined with the timestamp to form a new encrypted text. The addition of the timestamp ensures that the results of each encryption operation are different, which increases the security of the encrypted data. The combination process usually attaches the timestamp to the core data of the insurance application (such as the insurance application number, product code, etc.) to form a complete string or data structure as the encrypted input text.

[0190] The system concatenates or packages the insurance policy data information (such as insurance policy number, product code, policyholder name, etc.) with the acquired timestamp into a complete encrypted original text. The concatenation format can be:

[0191] "Insurance policy number - product code - policyholder name - ID number - timestamp". The combined data will be used as input for the encryption algorithm.

[0192] The combined original text is encrypted using the signature string as the key in combination with a 256-bit encryption algorithm (such as AES-256). AES-256 is a symmetric encryption technology with a key length of 256 bits, which can provide extremely high security. The signature string is derived from the MD5 hash operation or other generation mechanism, and it is used as a key to ensure that the encryption result is associated with the operator's identity. The system calls the encryption algorithm library (such as AES-256), uses the signature string as the key, and generates a fixed-length key input. Then, the system inputs the combined original text into the encryption function and performs the encryption operation. The encrypted data string generated in the end is ciphertext, and the original data cannot be read directly. The specific encryption process is as follows: encryptedData = AES-256 (encrypted original text, signature string).

[0193] In other specific embodiments, in order to further enhance the anti-replay capability of the data, the system not only uses timestamps, but also introduces a random number or a unique transaction ID combined with the insurance application data to form a dynamic encrypted original. This combination ensures that even if multiple encryptions are performed in a very short period of time, each encrypted data is unique. The system performs double-layer encryption on the original text. The first layer uses the operator's public key for asymmetric encryption, and the second layer uses the signature string as the AES-256 key for symmetric encryption. The encrypted data string generated by the double encryption prevents any unauthorized decryption and ensures high data security.

[0194] This embodiment combines the timestamp with the insurance application data information for encryption. Each encryption operation generates a unique encrypted data string to prevent security risks caused by repeated data encryption. At the same time, the encryption technology with a 256-bit key length ensures the security of data during transmission and storage, preventing unauthorized access or tampering. The signature string is used as an encryption key to bind the operator identity to the encryption process, enhancing the traceability of operations and the reliability of data protection.

[0195] The present invention also provides an insurance policy anti-counterfeiting device based on a two-dimensional code. Figure 2 , Figure 2 The following is a functional module diagram of a preferred embodiment of the insurance application form anti-counterfeiting device based on a QR code of the present invention. The insurance application form anti-counterfeiting device based on a QR code comprises:

[0196] The first-call task audio acquisition module is used to obtain the first-call task audio data of the claims adjuster;

[0197] An audio processing and segmentation module, used to identify and detect the silence interval in the audio data of the first connection task, and segment the audio data of the first connection task according to the length of the silence interval, so as to divide it into a plurality of independent voice segments;

[0198] A fuzzy speech library construction module is used to determine the standard speech keywords and construct a fuzzy speech library of the first standard speech based on the standard speech keywords;

[0199] The syllable comparison and execution evaluation module is used to compare the syllable units of the voice segment with the keyword syllables in the fuzzy voice library, and generate the anti-counterfeiting result of the insurance policy based on the QR code by the claims adjuster based on the syllable similarity.

[0200] The specific implementation of the QR code-based insurance form anti-counterfeiting device of the present invention is basically the same as the above-mentioned QR code-based insurance form anti-counterfeiting method embodiments, and will not be repeated here.

[0201] The present invention also provides an insurance policy anti-counterfeiting device based on a two-dimensional code, such as Figure 3 As shown, the insurance policy anti-counterfeiting device based on the QR code may include: a processor 1001, such as a CPU, a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. Among them, the communication bus 1002 is used to realize the connection and communication between these components. The user interface 1003 may include a display screen (Display), an input unit such as a keyboard (Keyboard), and the optional user interface 1003 may also include a standard wired interface and a wireless interface. The network interface 1004 may optionally include a standard wired interface and a wireless interface (such as a WI-FI interface). The memory 1005 may be a high-speed RAM memory, or a stable memory (non-volatile memory), such as a disk memory. The memory 1005 may also be a storage device independent of the aforementioned processor 1001.

[0202] Those skilled in the art will understand that Figure 3 The hardware structure of the insurance application anti-counterfeiting device based on the QR code does not constitute a limitation of the insurance application anti-counterfeiting device based on the QR code, and may include more or fewer components than shown in the figure, or a combination of certain components, or a different arrangement of components.

[0203] like Figure 3 As shown, the memory 1005 as a storage medium may include an operating system, a network communication module, a user interface module, and an insurance application form anti-counterfeiting program based on a QR code. The operating system is a program for managing and controlling the insurance application form anti-counterfeiting device and software resources based on a QR code, and supports the operation of the network communication module, the user interface module, the insurance application form anti-counterfeiting program based on a QR code, and other programs or software; the network communication module is used to manage and control the network interface 1004; the user interface module is used to manage and control the user interface 1003.

[0204] exist Figure 3 In the hardware structure of the QR code-based insurance policy anti-counterfeiting device shown, the network interface 1004 is mainly used to connect to the background server and communicate data with the background server; the user interface 1003 is mainly used to connect to the client and communicate data with the client; the processor 1001 can call the QR code-based insurance policy anti-counterfeiting program stored in the memory 1005, and perform the same operations as the QR code-based insurance policy anti-counterfeiting method.

[0205] The specific implementation of the QR code-based insurance form anti-counterfeiting device of the present invention is basically the same as the above-mentioned QR code-based insurance form anti-counterfeiting method embodiments, and will not be repeated here.

[0206] In addition, an embodiment of the present invention also proposes a computer storage medium, on which a QR code-based insurance form anti-counterfeiting program is stored. When the QR code-based insurance form anti-counterfeiting program is executed by a processor, the steps of the QR code-based insurance form anti-counterfeiting method as described above are implemented.

[0207] The specific implementation methods of the computer storage medium of the present invention are basically the same as the above-mentioned embodiments of the insurance policy anti-counterfeiting method based on QR code, and will not be repeated here.

[0208] It should be noted that if software tools or components other than those of our company appear in the embodiments of the present application, they are only used for illustrative purposes and do not represent actual use.

Claims

1. A QR code-based insurance application form anti-counterfeiting method, characterized in that: The following steps are involved: Obtaining insurance application data information, the insurance application data information including the insurance application number, product code, policyholder name and policyholder identification number; Obtain a unique identity identifier of the service operator, perform an MD5 hash operation on the unique identity identifier, and generate a signature string; The signature string is used as an encryption key, and the insurance application data information is used as the encrypted original text to perform an encryption operation to generate an encrypted data string; The encrypted data string is converted into two-dimensional code graphic data, and the two-dimensional code graphic data is embedded into the insurance application form to generate an insurance application form with an anti-counterfeiting two-dimensional code graphic.

2. The anti-counterfeiting method for insurance application form based on QR code according to claim 1, characterized in that: After the insurance application form with the anti-counterfeiting QR code is generated, it also includes: Obtaining an insurance application form to be verified, and extracting a data string to be verified from the anti-counterfeiting two-dimensional code graphic of the insurance application form to be verified; Decrypting the data string to be verified to restore the data information to be verified; Identify the insurance application number from the data information to be verified, and confirm the corresponding original insurance application data information based on the insurance application number; Compare the data information to be verified with the data information of the original insurance application form to verify whether the contents are consistent; If the contents are consistent, a pre-trained deep learning model is used to identify whether the target seal exists in the insurance application to be verified; If the target seal is identified, it is confirmed that the insurance application to be verified has passed the verification.

3. The anti-counterfeiting method for insurance application form based on QR code as claimed in claim 2, characterized in that: Identify whether the target seal exists in the insurance application form to be verified by using a pre-trained deep learning model, including: Extracting a seal image from the insurance application form to be verified, removing noise and performing binarization processing on the seal image to separate the seal area from the background; Inputting the processed seal image into a pre-trained deep learning model to identify the seal type in the seal image, wherein the seal type includes a company seal and a business-specific seal; According to the recognized seal type, compare it with the type of the target seal; If the seal type does not match, confirm that the target seal does not exist in the insurance application to be verified; If the seal type matches, extracting feature information corresponding to the identified seal type from the seal image, the feature information including the shape and texture of the seal; Matching the characteristic information with the characteristics of the target seal, and analyzing the similarity between the seal image and the target seal; If the similarity exceeds the similarity threshold, it is confirmed that the target seal exists in the insurance application to be verified.

4. The anti-counterfeiting method for insurance application form based on QR code as claimed in claim 2, characterized in that: After comparing the data information to be verified with the data information of the original insurance application form to verify whether the contents are consistent, the method further includes: If the contents are inconsistent, re-extract the data string to be verified from the anti-counterfeiting QR code graphic of the insurance application form to be verified; Check the structure of the data string to be verified to verify whether the data fields are complete and whether there are any missing fields; If there are missing or incomplete fields in the data string to be verified, extract redundant data from the anti-counterfeiting two-dimensional code image, the redundant data including a hash value, and correct the data string to be verified based on the redundant data; Compare the corrected data string to be verified with the original insurance application data information to verify again whether the contents are consistent; If the contents are consistent, a pre-trained deep learning model is used to identify whether the target seal exists in the insurance application to be verified; If the contents are inconsistent, the verification is confirmed to have failed, and a failure record is generated and sent to the administrator.

5. The anti-counterfeiting method for insurance application based on QR code according to claim 1, characterized in that: Converting the encrypted data string into two-dimensional code graphic data includes: Encoding the encrypted data string, converting the encrypted data string into encoded data in a two-dimensional code format according to a two-dimensional code encoding standard; Using a QR code generation tool to generate two-dimensional code graphic data according to the encoding data, the two-dimensional code graphic is composed of a two-dimensional matrix containing the encrypted data string; The two-dimensional code graphic data is output as an image format for embedding into the insurance application form.

6. The anti-counterfeiting method for insurance application form based on QR code according to claim 1, characterized in that: The two-dimensional code graphic data is embedded into the insurance application form to generate an insurance application form with an anti-counterfeiting two-dimensional code graphic, including: Confirm the type of the insurance application form, including electronic insurance application form and paper insurance application form; When the insurance application form is an electronic insurance application form, obtaining the resolution of the display device, and determining the size of the QR code graphic based on the resolution; Identify a blank area in the electronic insurance application page, and select a target blank area whose boundary exceeds a first preset value from the edge of the page and whose area meets the size requirement of the QR code image; Embedding the two-dimensional code graphic into an independent display layer at the position of the target blank area to generate an electronic insurance application form with an anti-counterfeiting two-dimensional code graphic; When the insurance application form is a paper insurance application form, the printing accuracy DPI of the printing device is obtained, and the size of the two-dimensional code graphic is determined based on the printing accuracy DPI; Identify a blank area in the paper insurance application page, and select a target blank area whose border is more than a second preset value from the edge of the page and whose area meets the size requirement of the QR code image; Embed the QR code image into the target blank area to generate a paper insurance policy with an anti-counterfeiting QR code graphic.

7. The anti-counterfeiting method for insurance application form based on QR code as claimed in claim 1, characterized in that: The signature string is used as an encryption key, and the insurance application data information is used as the encrypted original text, and an encryption operation is performed to generate an encrypted data string, including: Get the timestamp of the current encryption operation; Combining the insurance application data information with the timestamp as the encrypted original text; The signature string is used as an encryption key, and an encryption technology with a 256-bit key length is adopted to perform an encryption operation on the encrypted original text to generate an encrypted data string.

8. An anti-counterfeiting device for insurance application form based on QR code, characterized in that: The insurance policy anti-counterfeiting device based on the QR code includes: The data collection module is used to obtain the insurance application data information, wherein the insurance application data information includes the insurance application number, product code, name of the insured and the insured's identification number; An encryption key generation module is used to obtain a unique identity identifier of a service operator, perform an MD5 hash operation on the unique identity identifier, and generate a signature string; An encryption processing module, used to use the signature string as an encryption key and the insurance application data information as an encrypted original text to perform an encryption operation to generate an encrypted data string; The two-dimensional code generation and embedding module is used to convert the encrypted data string into two-dimensional code graphic data, embed the two-dimensional code graphic data into the insurance application form, and generate an insurance application form with an anti-counterfeiting two-dimensional code graphic.

9. An anti-counterfeiting device for insurance application based on a QR code, characterized in that: The QR code-based insurance form anti-counterfeiting device includes a memory, a processor, and a QR code-based insurance form anti-counterfeiting program stored in the memory and executable on the processor. When the QR code-based insurance form anti-counterfeiting program is executed by the processor, the steps of the QR code-based insurance form anti-counterfeiting method as described in any one of claims 1 to 7 are implemented.

10. A computer storage medium, characterized in that: The storage medium stores a QR code-based insurance application anti-counterfeiting program, which, when executed by the processor, implements the steps of the QR code-based insurance application anti-counterfeiting method as described in any one of claims 1 to 7.

Images