Rapid verification system for real-name DID and biological feature fusion

Abstract

The invention discloses a real-name DID and biological feature fusion rapid verification system, and the system comprises the following modules: a biological feature processing module which is used for generating a nested biological feature vector; the hash generation module is used for performing hash operation and constructing a Merkle tree; the real-name verification module is used for generating real-name identity data; the DID construction module is used for generating a DID document and adding a BioBinding extension field; the proof generation module is used for generating a verification record; the block chain writing module is used for forming an on-chain real-name DID record; the verification module is used for matching with the on-chain real-name DID record; the confidence control module is used for comparing the matching result with a preset verification threshold value, and if verification is passed, authorized access is completed; and if the verification is not passed, triggering a multi-factor re-verification mechanism to perform secondary verification, calculating a DID confidence value, and if the secondary verification is still not passed, executing identity freezing. According to the method, multi-modal biological recognition and a real-name DID generation mechanism are fused, and on-chain credible verification is realized.

Description

Technical Field

[0001] The present invention relates to the field of digital identity recognition and data privacy technology, and in particular to a rapid verification system that integrates real-name DID and biometrics. Background Art

[0002] With the widespread development of information technology, digital identity authentication has become a core component of cyberspace security. Especially in high-security scenarios such as financial services, government travel, healthcare, and social platforms, identity recognition systems often need to ensure efficient identification while also strengthening user privacy and data credibility. However, traditional username-password authentication methods, due to vulnerabilities such as credential theft, high risk of replay attacks, and irreversibility, have been gradually replaced by more secure biometric-based authentication systems.

[0003] Currently, more mature biometric technologies include facial recognition, fingerprint recognition, voiceprint recognition, iris recognition, etc. Although these single-modal recognition methods can improve verification security, they still have many problems in practical applications. For example, a single modality is limited by the accuracy of the acquisition equipment and environmental interference, and its recognition accuracy is easily affected by factors such as image blur, noise pollution, or counterfeit attacks. To this end, in recent years, researchers have begun to try to introduce multimodal fusion recognition technology to improve recognition robustness by integrating information from multiple biological channels. However, existing multimodal fusion solutions mostly remain at the level of shallow feature splicing or simple weighting, lacking effective deep nested structure modeling methods, making it difficult to fully explore the potential structural coupling relationship between multiple modalities, resulting in limited fusion effects.

[0004] On the other hand, with the rise of blockchain technology, decentralized identity (DID) has gradually become a key path for the next generation of digital identity management. DID is generated and controlled by users, and features autonomy, strong privacy protection, and high verifiability. It has been recognized as a core standard by the W3C organization. However, most current DID implementations still use static public keys or user-declared attributes for identity binding, lacking a mechanism for deep integration with users' real physical characteristics (especially biometrics in high-security scenarios). Because DID itself does not have real-name attributes, without a trusted binding mechanism, even if the DID on-chain records are complete, it is difficult to effectively trace the authenticity of the user, which cannot meet the stringent requirements for trusted on-chain identities in government, finance, and judicial scenarios.

[0005] Some existing academic research and industrial practices attempt to improve the real-person credibility of the DID system by writing biometric hashes onto the chain and establishing a weak binding relationship with the DID. However, due to the limited accuracy of the biometric hash algorithm and the high cost of on-chain storage, most implementations only use compressed single-modal hash values on the chain, lacking the ability to express modal fusion and lacking dynamic update and identity status backtracking mechanisms. As a result, once the identity chain is established, it cannot be adjusted or revoked. In addition, there is currently a lack of technical paths to achieve high consistency expansion of multimodal fusion biometrics and DID structure specifications (such as the W3C DID document structure), resulting in a chaotic on-chain identity record structure and unfavorable system parsing and verification.

[0006] In terms of security verification, most existing on-chain identity verification schemes are based on plaintext comparison, fingerprint duplication checking, or feature signature comparison, and lack a verifiable logical expression of the binding relationship between identity and biometric hash. Although some schemes based on multi-party secure computing (MPC) or homomorphic encryption have privacy protection capabilities, they have high computational overhead and complex on-chain deployment, making them unsuitable for identity verification applications with high real-time requirements. Zero-knowledge proof (ZKP), an important direction in the field of privacy computing in recent years, is very suitable for private verification in the identity binding process due to its ability to verify facts without exposing the original data. However, the ZKP integration mechanism for binding with biometrics has not yet been widely implemented in existing DID systems, and there is a lack of complete protocol implementation and structural definition.

[0007] Furthermore, in actual identity verification, user status is often not static. Device changes, environmental drift, and changes in behavioral characteristics can all affect the credibility of the identification system's judgment of the current identity. However, traditional identity authentication systems often lack the ability to model users' dynamic behavioral states and cannot automatically adjust verification intensity based on risk levels. For example, a user may change from a trusted device to an unfamiliar terminal at a certain point in time. If the system cannot detect and intervene in time, it may allow a forged identity to pass authentication. Conversely, if multi-factor re-verification can be triggered when confidence levels fluctuate abnormally, and the identity credibility score is dynamically adjusted based on a sliding window mechanism, the risk can be effectively reduced. However, current DID systems generally do not integrate this type of dynamic confidence control capability.

[0008] Therefore, how to provide a fast verification system that integrates real-name DID and biometrics is an urgent problem that technicians in this field need to solve. Summary of the Invention

[0009] One purpose of the present invention is to propose a rapid verification system that integrates real-name DID and biometric features. The present invention integrates multimodal biometrics with real-name DID generation mechanisms, constructs nested feature coding and BioBinding extension structures, and combines zero-knowledge proof with on-chain matching to achieve a highly reliable, low-leakage, and efficient identity verification process. It has the advantages of strong privacy protection, high verification accuracy, fast response speed, and scalable structure, and is suitable for high-security scenarios such as government affairs, finance, and entry and exit.

[0010] The rapid verification system integrating real-name DID and biometric features according to an embodiment of the present invention includes the following modules:

[0011] A biometric processing module, which is used to collect multimodal biometric data of users and generate nested biometric vectors through a variational nested encoder;

[0012] A hash generation module, configured to perform a hash operation on the nested biometric feature vector and construct a Merkle tree to generate a biometric hash root value;

[0013] Real-name verification module, used to verify the authenticity of the user's real-name information and generate real-name identity data;

[0014] A DID building module is used to generate a DID document that complies with the W3C standard format and adds a BioBinding extension field to record the on-chain reference address of the biohash root value;

[0015] a proof generation module for constructing a zero-knowledge proof based on the real-name identity data and the biometric hash root value, and generating a verification record;

[0016] A blockchain writing module is used to write the DID document, BioBinding extension field, and verification record into the blockchain to form an on-chain real-name DID record;

[0017] The verification module is used to collect user's single-modal lightweight biometric data during identity verification, generate a summary vector, and match it with the real-name DID record on the chain;

[0018] The confidence control module is used to compare the matching results with the preset verification threshold. If the verification passes, the access is authorized. If the verification fails, the multi-factor re-verification mechanism is triggered for a second verification and the DID confidence value is calculated. If the second verification still fails, the identity is frozen.

[0019] The rapid verification method for integrating real-name DID and biometric features according to an embodiment of the present invention includes the following steps:

[0020] S1. Collect the user's multimodal biometric data and input it into the variational nested encoder model to fuse and compress the mapping to generate a nested biometric vector;

[0021] S2. Performing a hash operation on the nested biometric feature vector and constructing a Merkle tree to generate a biometric hash root value;

[0022] S3. Obtain the real-name information provided by the user, verify the authenticity of the real-name information, and generate real-name identity data;

[0023] S4. Generate a DID document in accordance with the W3C standard format based on the real-name identity data, and add a BioBinding extension field to record the on-chain reference address of the biohash root value;

[0024] S5. Construct a zero-knowledge proof protocol input, use the zero-knowledge proof protocol to verify the binding relationship between the real-name identity data and the biometric hash root value, and generate a verification record;

[0025] S6. Write the DID document, BioBinding extension field, and verification record into the blockchain to generate an immutable on-chain real-name DID record.

[0026] S7. During identity verification, collect the user's single-modal lightweight biometric data, locally process the single-modal lightweight biometric data to generate a summary vector, and submit it to the on-chain comparison engine for matching with the on-chain real-name DID record;

[0027] S8. Compare the matching result with the preset verification threshold. If the verification passes, authorized access is completed; if the verification fails, the multi-factor re-verification mechanism is triggered for a second verification and the DID confidence value is calculated. If the second verification still fails, the identity is frozen.

[0028] Optionally, the multimodal biometric data includes a facial image, a fingerprint image, and a voiceprint signal.

[0029] Optionally, the S1 specifically includes:

[0030] S11, the variational nested encoder model includes a modal feature extraction layer, a cross-modal fusion layer, a nested variational encoder layer and a compression mapping output layer;

[0031] S12, inputting the multimodal biometric data into the modal feature extraction layer, performing convolutional coding, normalization, and activation processing on each modality, and extracting preliminary modal features;

[0032] S13. Inputting the preliminary modal features into the cross-modal fusion layer, wherein the cross-modal fusion layer includes a multi-head attention mechanism and a gating mechanism for dynamically adjusting the weight relationship between the modalities to generate a modal fusion representation;

[0033] S14. Input the modality fusion representation into the nested variational encoder layer, where the nested variational encoder layer consists of a main encoder and three sub-encoders, wherein the main encoder uses a multi-layer perceptron structure to model the modality fusion representation to extract the global potential mean and variance vector shared by all modalities;

[0034] S15, the three sub-encoders correspond to facial images, fingerprint images, and voiceprint signals, respectively, use independent variational channels in the shared architecture to model local variation information of their respective modalities, and output modality-specific latent vectors;

[0035] S16, fusing the global latent mean and variance vector output by the main encoder with the modal local latent vector output by each sub-encoder to generate a nested latent feature vector;

[0036] S17: Input the nested latent feature vector into the compression mapping output layer, and obtain a structure-preserving nested biological feature vector through nonlinear dimensionality reduction mapping.

[0037] Optionally, the S2 specifically includes:

[0038] S21. Structurally divide the nested biometric feature vector into a number of sub-vector segments according to modality source and feature level, where each sub-vector segment corresponds to a feature set in a biometric modality or encoding level, for use in distributed hash calculations;

[0039] S22. Perform three rounds of hashing operations on each subvector fragment using the SHA-3 hash algorithm. The first round performs hashing after vector normalization, the second round performs hashing after vector dimension rotation, and the third round performs scrambled hashing after introducing a mask parameter. Multiple hash subvalues are generated as leaf node values of the Merkle tree.

[0040] S23. Construct the underlying structure of the Merkle tree. Use a sequential pairing method to concatenate two adjacent hash sub-values and then execute the SHA-3 hash algorithm again to generate the corresponding parent node hash value. Iterate upward layer by layer to generate intermediate nodes until the root node is generated.

[0041] S24. During the construction of each layer of the Merkle tree, a modality identification tag is attached to each intermediate node. The modality identification tag represents the modality source combination contained in the intermediate node through a coding mapping method, which is used to trace the modality composition relationship;

[0042] S25. A local update mechanism is introduced during the construction process. The input hash combination of each intermediate node is stored in a local buffer and an index table is constructed. When a subvector fragment changes, only the corresponding branch and upper-level node are updated to avoid repeated construction of the entire Merkle tree.

[0043] S26. After completing the calculation of all intermediate nodes, the biometric hash root value is finally output as the unique identification hash root value of the nested biometric feature vector.

[0044] Optionally, the S4 specifically includes:

[0045] S41. Construct a basic identity information object based on the real-name identity data. The basic identity information object includes a user identifier, an issuing institution identifier, an issuance timestamp, an expiration timestamp, and an identity identification code, and is organized in a key-value pair format.

[0046] S42. Generate a DID document according to the W3C Decentralized Identity Technical Specification. The DID document includes a context field, an id field, an authentication field, and a verificationMethod field. The context field specifies the specification version, and the id field corresponds to the user's DID identifier.

[0047] S43. Set the elliptic curve-based public key identification data in the verificationMethod field, and set the signature method identifier for identity authentication in the authentication field;

[0048] S44. Add an extended field BioBinding to the DID document. The BioBinding field includes four subfields: bioMerkleRoot, bindingTime, dataType, and storageReference, which correspond to the biometric hash root value, binding time, feature data type, and on-chain storage address information of the nested biometric feature vector, respectively.

[0049] S45. Encapsulate the DID document containing the BioBinding extension field into JSON-LD format.

[0050] Optionally, the S5 specifically includes:

[0051] S51. Construct an input data set for the zero-knowledge proof protocol, and represent the real-name identity data and the biometric hash root value as the private input of the relationship to be proved;

[0052] S52. Introduce zero-knowledge proof protocol verification, define constraint conditions based on the relationship to be proved, and express the constraint conditions in the form of a circuit, wherein the circuit form includes the constraint logic of the relationship between the real-name identity data and the biometric hash root value;

[0053] S53. Generate a zero-knowledge proof public-private key pair based on the constraints of the circuit form, where the private key is used to generate the proof and the public key is used for on-chain verification;

[0054] S54: Use the private key to perform zero-knowledge proof on the binding relationship between the real-name identity data and the biometric hash root value, and generate proof data, wherein the proof data includes a proof subject, a random challenge number, and a response factor;

[0055] S55. Structured encoding of the certification data is performed, and digital signature is performed;

[0056] S56. Combine the digitally signed proof data with the corresponding public key to form a verification record as a verification record of the binding relationship between the real-name identity data and the biometric hash root value. The verification record includes a signature summary and timestamp information.

[0057] Optionally, when writing the DID document, BioBinding extension field and verification record to the blockchain, the pre-deployed smart contract registration interface is called, and the real-name DID record on the chain is used as the transaction payload and recorded in the block event log.

[0058] Optionally, the S7 specifically includes:

[0059] S71. During identity verification, collect single-modality lightweight biometric data of the user in real time through the terminal device, and perform scale normalization, noise filtering, and standardization on the single-modality lightweight biometric data to obtain a feature preprocessing vector.

[0060] S72. A fixed random projection matrix and a bias vector are stored locally on the terminal device, wherein the random projection matrix is a predefined random value matrix, and the bias vector is a predefined random constant vector. A summary vector is generated by performing a nonlinear activation function ReLU mapping on the feature preprocessing vector.

[0061] S73. For each element of the summary vector, perform binary quantization dimension by dimension using the sign function of the local-aware hash algorithm, convert each element into a corresponding binary sign bit, and form a local-aware hash string with 64 bits:

[0062]

[0063] Among them, h i Represents the value of the i-th bit in the local perception hash string, si represents the i-th element in the summary vector;

[0064] S74. Establish a communication connection with the blockchain network through the terminal device, encapsulate the local perception hash string as an on-chain transaction payload, and submit it to the on-chain comparison engine;

[0065] S75. The on-chain comparison engine reads the on-chain real-name DID record, parses the DID document to verify the validity of the source, extracts the biometric hash root value in the BioBinding extension field, and verifies the signature summary and timestamp information in the record, and receives the local perception hash string and matches it with the biometric hash root value;

[0066] S76. The on-chain comparison engine uses the Hamming distance function to calculate the similarity d between the local perception hash string and the biological hash root value. H :

[0067]

[0068] Among them, H bio,i Indicates the value of the i-th bit of the biometric hash root value, and d H Output as matching results.

[0069] d H Compare with the preset verification threshold θ, when d H When ≤θ, it means the verification is passed. H >θ, indicating that the verification failed.

[0070] Optionally, the multi-factor re-verification mechanism specifically includes calculating the DID confidence value based on the user device type, historical login geographic location, access frequency and verification behavior trajectory, and calculating the DID confidence change rate based on a sliding time window. When the confidence change rate exceeds the set confidence threshold, it is determined that the secondary verification still fails and the identity is frozen.

[0071] The beneficial effects of the present invention are:

[0072] First, this invention introduces a variational nested encoder model to perform a deep, structured fusion of multimodal biometrics, including facial images, fingerprint images, and voiceprint signals. It then extracts global latent semantics and modal local variation information through a master-sub-nested modeling approach, significantly improving the expressive integrity and discriminative power of the fused biometrics. Compared to traditional shallow splicing or simple weighted fusion methods, this invention achieves deep collaboration and unified representation between modalities, providing high-quality, low-redundancy feature input for subsequent identity binding.

[0073] Secondly, the present invention constructs an improved multimodal hashing mechanism based on Merkle trees. Through modal structure identification and a local update strategy, it implements an efficient and traceable nested hashing process and generates a unique biometric hash root value for identity binding. This hashing mechanism not only improves the accuracy of fuzzy matching but also significantly enhances the structural expressiveness and dynamic adaptability of biometric data during on-chain recording, resolving the issues of hashing being non-scalable and non-reconfigurable in existing systems.

[0074] In addition, the present invention proposes a standardized extension method for the DID document structure, introduces the BioBinding extension field based on the W3C specification, and records the biometric hash root value, binding time, biometric data type and on-chain storage address through a clear field structure, thereby achieving a strong binding of real-name identity data and multimodal biometric hash, and has the technical capabilities of on-chain structure parsability and field verification, filling the gap in the existing DID system's insufficient native support for biometric binding information.

[0075] Furthermore, this invention incorporates a zero-knowledge proof protocol to achieve verifiable modeling of the binding relationship between real-name identity data and biometric hash root values without exposing the original identity data and biometric features. This mechanism expresses the binding logic through circuit constraints, generates a verifiable proof structure, and binds it to the DID document to form a verification record. This meets the real-name verification requirements of high privacy, high trust, and high compliance. It is technically superior to traditional MPC or hash signature schemes and has greater practicality and scalability.

[0076] Finally, in the verification phase, the present invention designs a complete process that supports local perception hashing, on-chain fuzzy comparison, and dynamic confidence control. During the user identity verification process, the system collects lightweight biometric features of a single modality, generates a summary hash vector through random projection and symbol mapping, and performs Hamming distance calculation with the biometric hash root value in the on-chain real-name DID record to ensure a balance between matching efficiency and error tolerance. At the same time, the present invention constructs an identity situation map based on multi-source behavioral features such as user device, location, frequency, and introduces a sliding window mechanism to calculate the DID confidence change rate in real time, so that the identity system can automatically trigger multi-factor re-verification or perform identity freezing when encountering abnormal behavior, effectively improving the risk response capability and dynamic credibility of the overall verification system. BRIEF DESCRIPTION OF THE DRAWINGS

[0077] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0078] Figure 1 This is a schematic diagram of the structure of the rapid verification system that integrates real-name DID and biometric features proposed in the present invention;

[0079] Figure 2 This is the overall flow chart of the rapid verification method proposed in the present invention that integrates real-name DID and biometric features. DETAILED DESCRIPTION

[0080] The present invention will now be described in further detail with reference to the accompanying drawings, which are simplified schematic diagrams that illustrate the basic structure of the present invention in a schematic manner.

[0081] refer to Figure 1 The rapid verification system that integrates real-name DID and biometrics includes the following modules:

[0082] A biometric processing module, which is used to collect multimodal biometric data of users and generate nested biometric vectors through a variational nested encoder;

[0083] A hash generation module, configured to perform a hash operation on the nested biometric feature vector and construct a Merkle tree to generate a biometric hash root value;

[0084] Real-name verification module, used to verify the authenticity of the user's real-name information and generate real-name identity data;

[0085] A DID building module is used to generate a DID document that complies with the W3C standard format and adds a BioBinding extension field to record the on-chain reference address of the biohash root value;

[0086] a proof generation module for constructing a zero-knowledge proof based on the real-name identity data and the biometric hash root value, and generating a verification record;

[0087] A blockchain writing module is used to write the DID document, BioBinding extension field, and verification record into the blockchain to form an on-chain real-name DID record;

[0088] The verification module is used to collect user's single-modal lightweight biometric data during identity verification, generate a summary vector, and match it with the real-name DID record on the chain;

[0089] The confidence control module is used to compare the matching results with the preset verification threshold. If the verification passes, the access is authorized. If the verification fails, the multi-factor re-verification mechanism is triggered for a second verification and the DID confidence value is calculated. If the second verification still fails, the identity is frozen.

[0090] The present invention constructs a real-name DID rapid verification system that includes modules such as biometric processing, hash generation, real-name verification, DID construction, proof generation, on-chain writing, verification processing and confidence control, thus completely opening up a trusted fusion path between user biometrics and decentralized identities on the chain. The system design has a clear structure and clear module boundaries, and can achieve a full-process closed loop from data collection, fusion coding, identity generation, privacy verification to on-chain comparison and confidence control. Compared with traditional DID systems that are only bound based on static public keys or declared data, this system enhances the real-person credibility of DID and the on-chain verification response capability through structured multimodal biometric input and dynamic security control mechanisms. It has higher security, flexibility and scalability, and is suitable for high-security scenarios such as financial authentication, government real-name, and entry and exit control.

[0091] refer to Figure 2 The rapid verification method integrating real-name DID and biometric features includes the following steps:

[0092] S1. Collect the user's multimodal biometric data and input it into the variational nested encoder model to fuse and compress the mapping to generate a nested biometric vector;

[0093] S2. Performing a hash operation on the nested biometric feature vector and constructing a Merkle tree to generate a biometric hash root value;

[0094] S3. Obtain the real-name information provided by the user, verify the authenticity of the real-name information, and generate real-name identity data;

[0095] S4. Generate a DID document in accordance with the W3C standard format based on the real-name identity data, and add a BioBinding extension field to record the on-chain reference address of the biohash root value;

[0096] S5. Construct a zero-knowledge proof protocol input, use the zero-knowledge proof protocol to verify the binding relationship between the real-name identity data and the biometric hash root value, and generate a verification record;

[0097] S6. Write the DID document, BioBinding extension field, and verification record into the blockchain to generate an immutable on-chain real-name DID record.

[0098] S7. During identity verification, collect the user's single-modal lightweight biometric data, locally process the single-modal lightweight biometric data to generate a summary vector, and submit it to the on-chain comparison engine for matching with the on-chain real-name DID record;

[0099] S8. Compare the matching result with the preset verification threshold. If the verification passes, authorized access is completed; if the verification fails, the multi-factor re-verification mechanism is triggered for a second verification and the DID confidence value is calculated. If the second verification still fails, the identity is frozen.

[0100] This method defines the interaction method and sequence between the modules in the system, and clarifies the operational process of the entire real-name DID generation and verification process. Through the sequential processing method of S1 to S8, the system can first fuse multimodal biometrics to generate a unique identification hash value, then write it to the blockchain after zero-knowledge binding with the real-name data, and perform rapid comparison and dynamic control in the subsequent verification stage. This process design not only ensures the consistency between data structures, but also realizes a closed-loop link through upstream and downstream data dependencies, effectively improving verification efficiency and identity credibility. This process avoids the problems of module fragmentation, inconsistent data sources, and unverifiable verification links in existing technologies, and enhances the operability and consistency of the system in real application scenarios.

[0101] In this embodiment, the multimodal biometric data includes facial images, fingerprint images, and voiceprint signals. This modal combination can provide complementary identification information in different environments. For example, in scenarios where image acquisition is limited, fingerprint or voiceprint recognition can be switched to improve the stability and accuracy of verification. At the same time, multimodal input can effectively reduce the risk of counterfeit attacks and prevent security risks caused by the forgery of a single modality. Compared with the single-modality recognition solution in existing systems, this method has the advantages of strong anti-interference ability, wide feature coverage, and high authentication robustness. It is suitable for achieving secure identity verification in complex scenarios such as high-risk, cross-device, and multi-terminal scenarios.

[0102] In this embodiment, S1 specifically includes:

[0103] S11, the variational nested encoder model includes a modal feature extraction layer, a cross-modal fusion layer, a nested variational encoder layer and a compression mapping output layer;

[0104] S12, inputting the multimodal biometric data into the modal feature extraction layer, performing convolutional coding, normalization, and activation processing on each modality, and extracting preliminary modal features;

[0105] S13. Inputting the preliminary modal features into the cross-modal fusion layer, wherein the cross-modal fusion layer includes a multi-head attention mechanism and a gating mechanism for dynamically adjusting the weight relationship between the modalities to generate a modal fusion representation;

[0106] S14. Input the modality fusion representation into the nested variational encoder layer, where the nested variational encoder layer consists of a main encoder and three sub-encoders, wherein the main encoder uses a multi-layer perceptron structure to model the modality fusion representation to extract the global potential mean and variance vector shared by all modalities;

[0107] S15, the three sub-encoders correspond to facial images, fingerprint images, and voiceprint signals, respectively, use independent variational channels in the shared architecture to model local variation information of their respective modalities, and output modality-specific latent vectors;

[0108] S16, fusing the global latent mean and variance vector output by the main encoder with the modal local latent vector output by each sub-encoder to generate a nested latent feature vector;

[0109] S17: Input the nested latent feature vector into the compression mapping output layer, and obtain a structure-preserving nested biological feature vector through nonlinear dimensionality reduction mapping.

[0110] Step S1 defines the structural composition of the variational nested encoder in a hierarchical manner, including a modal feature extraction layer, a cross-modal fusion layer, a nested variational encoder layer, and a compression mapping output layer. The structural design ensures the hierarchical, interactive, and compact expression of multimodal features. This structure can model the feature distribution within the modality and the collaborative relationship between modalities, respectively, and extract global and local semantics through a master-sub nested mechanism, ultimately outputting a nested biometric feature vector that is unique and compressible. Compared to traditional feature splicing encoding methods, this approach has stronger discriminability and generation consistency, can significantly improve the accuracy and efficiency of subsequent hash matching and identity binding, and is an effective innovation in the structural expression of current biometric technology in decentralized identity systems.

[0111] In this embodiment, S2 specifically includes:

[0112] S21. Structurally divide the nested biometric feature vector into a number of sub-vector segments according to modality source and feature level, where each sub-vector segment corresponds to a feature set in a biometric modality or encoding level, for use in distributed hash calculations;

[0113] S22. Perform three rounds of hashing operations on each subvector fragment using the SHA-3 hash algorithm. The first round performs hashing after vector normalization, the second round performs hashing after vector dimension rotation, and the third round performs scrambled hashing after introducing a mask parameter. Multiple hash subvalues are generated as leaf node values of the Merkle tree.

[0114] S23. Construct the underlying structure of the Merkle tree. Use a sequential pairing method to concatenate two adjacent hash sub-values and then execute the SHA-3 hash algorithm again to generate the corresponding parent node hash value. Iterate upward layer by layer to generate intermediate nodes until the root node is generated.

[0115] S24. During the construction of each layer of the Merkle tree, a modality identification tag is attached to each intermediate node. The modality identification tag represents the modality source combination contained in the intermediate node through a coding mapping method, which is used to trace the modality composition relationship;

[0116] S25. A local update mechanism is introduced during the construction process. The input hash combination of each intermediate node is stored in a local buffer and an index table is constructed. When a subvector fragment changes, only the corresponding branch and upper-level node are updated to avoid repeated construction of the entire Merkle tree.

[0117] S26. After completing the calculation of all intermediate nodes, the biometric hash root value is finally output as the unique identification hash root value of the nested biometric feature vector.

[0118] Step S2 proposes an improved process for structural partitioning, multi-round scrambled hashing, and Merkle tree construction for nested biometric feature vectors, forming a nested hash structure with modal traceability and an update mechanism. Through the modal and hierarchical partitioning mechanism, the introduction of mask perturbations, and the identification of node labels in the Merkle tree, not only is the collision resistance and security of the biometric hash improved, but the collaborative expression of feature modalities and structures is also achieved in the hash chain construction. In particular, the addition of a local update mechanism during the construction process allows only the corresponding node to be updated when a certain modality changes, avoiding full tree reconstruction and significantly improving the system's real-time performance and computational efficiency. This solution improves the uniqueness and adaptability of on-chain bound data and is suitable for scenarios where multimodal features frequently change.

[0119] In this embodiment, the S4 specifically includes:

[0120] S41. Construct a basic identity information object based on the real-name identity data. The basic identity information object includes a user identifier, an issuing institution identifier, an issuance timestamp, an expiration timestamp, and an identity identification code, and is organized in a key-value pair format.

[0121] S42. Generate a DID document according to the W3C Decentralized Identity Technical Specification. The DID document includes a context field, an id field, an authentication field, and a verificationMethod field. The context field specifies the specification version, and the id field corresponds to the user's DID identifier.

[0122] S43. Set the elliptic curve-based public key identification data in the verificationMethod field, and set the signature method identifier for identity authentication in the authentication field;

[0123] S44. Add an extended field BioBinding to the DID document. The BioBinding field includes four subfields: bioMerkleRoot, bindingTime, dataType, and storageReference, which correspond to the biometric hash root value, binding time, feature data type, and on-chain storage address information of the nested biometric feature vector, respectively.

[0124] S45. Encapsulate the DID document containing the BioBinding extension field into JSON-LD format.

[0125] Step S4 implements a standardized extension to the DID document structure. Without compromising the integrity of the W3C standard structure, a new BioBinding extension field is added to record the biometric hash root value and its associated on-chain reference information. Through the refined design of fields such as bioMerkleRoot, bindingTime, dataType, and storageReference, modular expansion and field-level verifiability of the on-chain identity structure are achieved. This extension approach not only provides excellent system compatibility but also improves DID document parsing efficiency and on-chain structure readability. It effectively addresses the inability of traditional DID systems to carry structured biometric information, providing identity systems with more complete and auditable on-chain expression capabilities.

[0126] In this embodiment, the S5 specifically includes:

[0127] S51. Construct an input data set for the zero-knowledge proof protocol, and represent the real-name identity data and the biometric hash root value as the private input of the relationship to be proved;

[0128] S52. Introduce zero-knowledge proof protocol verification, define constraint conditions based on the relationship to be proved, and express the constraint conditions in the form of a circuit, wherein the circuit form includes the constraint logic of the relationship between the real-name identity data and the biometric hash root value;

[0129] S53. Generate a zero-knowledge proof public-private key pair based on the constraints of the circuit form, where the private key is used to generate the proof and the public key is used for on-chain verification;

[0130] S54: Use the private key to perform zero-knowledge proof on the binding relationship between the real-name identity data and the biometric hash root value, and generate proof data, wherein the proof data includes a proof subject, a random challenge number, and a response factor;

[0131] S55. Structured encoding of the certification data is performed, and digital signature is performed;

[0132] S56. Combine the digitally signed proof data with the corresponding public key to form a verification record as a verification record of the binding relationship between the real-name identity data and the biometric hash root value. The verification record includes a signature summary and timestamp information.

[0133] Step S5 implements a zero-knowledge proof protocol to privately verify the binding relationship between real-name identity data and the biometric hash root value, establishing an identity association mechanism that balances authenticity and privacy. Through circuit-level constraint modeling, key generation, proof construction, and signature encapsulation, a verifiable proof structure is established that verifies the binding relationship without leaking the original data. This solution not only reduces the risk of sensitive information being exposed on-chain but also strengthens the auditability and non-repudiation capabilities of the binding relationship. Compared to existing plaintext signature or static binding methods, it offers a higher level of security and data protection, making it suitable for identity authentication scenarios with high privacy and compliance requirements.

[0134] In this embodiment, when the DID document, BioBinding extension field and verification record are written to the blockchain, the pre-deployed smart contract registration interface is called, and the real-name DID record on the chain is used as the transaction payload and recorded in the block event log.

[0135] By writing the DID document, BioBinding extension fields, and verification records as transaction payloads to the blockchain and recording them in the block event log, we achieve on-chain, trusted proof of ownership of the real-name identity structure. This mechanism not only ensures data immutability and full-chain traceability, but also standardizes the data writing process through contract interfaces, improving system consistency and compatibility. Compared to traditional off-chain record storage or centralized database storage methods, this solution enhances the long-term credibility and cross-party sharing capabilities of identity data, and is a key technical path to achieving decentralized real-name identity governance.

[0136] In this embodiment, the S7 specifically includes:

[0137] S71. During identity verification, collect single-modality lightweight biometric data of the user in real time through the terminal device, and perform scale normalization, noise filtering, and standardization on the single-modality lightweight biometric data to obtain a feature preprocessing vector.

[0138] S72. A fixed random projection matrix and a bias vector are stored locally on the terminal device, wherein the random projection matrix is a predefined random value matrix, and the bias vector is a predefined random constant vector. A summary vector is generated by performing a nonlinear activation function ReLU mapping on the feature preprocessing vector.

[0139] S73. For each element of the summary vector, perform binary quantization dimension by dimension using the sign function of the local-aware hash algorithm, convert each element into a corresponding binary sign bit, and form a local-aware hash string with 64 bits:

[0140]

[0141] Among them, h i Represents the value of the i-th bit in the local perception hash string, s i represents the i-th element in the summary vector;

[0142] S74. Establish a communication connection with the blockchain network through the terminal device, encapsulate the local perception hash string as an on-chain transaction payload, and submit it to the on-chain comparison engine;

[0143] S75. The on-chain comparison engine reads the on-chain real-name DID record, parses the DID document to verify the validity of the source, extracts the biometric hash root value in the BioBinding extension field, and verifies the signature summary and timestamp information in the record, and receives the local perception hash string and matches it with the biometric hash root value;

[0144] S76. The on-chain comparison engine uses the Hamming distance function to calculate the similarity d between the local perception hash string and the biological hash root value. H :

[0145]

[0146] Among them, H bio,i Indicates the value of the i-th bit of the biometric hash root value, and d H Output as matching results.

[0147] d H Compare with the preset verification threshold θ, when d H When ≤θ, it means the verification is passed. H >θ, indicating that the verification failed.

[0148] Step S7 proposes a rapid verification mechanism based on lightweight feature extraction and fuzzy hash matching, significantly improving the system's response speed and fault tolerance. Feature vectors are mapped to 64-bit hash strings using a local-aware hash function, and similarity is assessed by calculating the Hamming distance with the on-chain hash value. This enables a lightweight comparison process without uploading the original image. This method ensures privacy while supporting approximate matching and fault tolerance, effectively resolving existing recognition failures caused by biometric collection errors. It is suitable for applications requiring high real-time performance, such as airport security, access control, and remote account opening.

[0149] In this embodiment, the multi-factor re-verification mechanism specifically includes calculating the DID confidence value based on the user's device type, historical login location, access frequency, and verification behavior trajectory, and calculating the DID confidence change rate based on a sliding time window. When the confidence change rate exceeds the set confidence threshold, it is determined that the secondary verification has still failed, and the identity is frozen. By introducing a dynamic calculation mechanism for the DID confidence value and change rate, risk perception and behavioral trend tracking of the user's identity status are achieved. By combining multi-dimensional behavioral data such as user device type, location trajectory, and access frequency to construct an identity situation map, and judging the confidence fluctuation trend based on a sliding time window, the system can trigger multi-factor re-verification when the risk is high and automatically freeze the identity after verification fails. This mechanism greatly improves the system's responsiveness to abnormal behavior, has the advantages of dynamic control, self-learning, and state recovery, and is a key component for realizing the intelligent and defensive upgrade of the trusted identity management system.

[0150] Example 1:

[0151] In order to verify the feasibility of the present invention in implementation, the present invention is applied to the intelligent identity verification system of a large-scale urban comprehensive government service center. The service center receives more than 3,000 users on average every day, covering multiple high-security business windows such as identity registration, household registration changes, social security processing, real estate registration, etc. The original system mainly relies on users to present physical identification documents, and uses the ID card reader to compare the identity with the on-site face recognition module. However, in actual operation, problems such as document loss, low equipment recognition rate, and large light interference frequently occur. Especially in the case of multiple people concurrently, high environmental noise or users wearing masks, the system misrecognition rate is high, manual intervention is frequent, and the user experience is poor. The center urgently needs a more efficient, secure, and privacy-friendly identity authentication solution that can improve traffic efficiency and system stability while ensuring real-name compliance requirements.

[0152] The present invention deploys a rapid verification system that integrates real-name DID and multimodal biometrics at the verification entrance and business front desk of the center. When a user accesses the system for the first time, the on-site guide assists the user in entering his or her facial image, fingerprint image and voiceprint data through the self-service terminal, and uses the real-name information accessed by the public security port for simultaneous verification. After the acquisition is completed, the system uses the variational nested encoder in the present invention to fuse the three-modal features, and constructs a Merkle hash tree to generate a biometric hash root value, and then generates a DID document containing the BioBinding extension field. The user's real-name information and the biometric hash relationship are verified through a zero-knowledge proof protocol, and the complete DID structure is written to the blockchain. The entire process performs modeling and compression off-chain, and only uploads the structure hash value and verification record to the chain to ensure that private data does not exit the terminal and does not remain on the chain.

[0153] When the user visits the center again or performs online identity authentication in the future, he only needs to collect lightweight biometric information of any modality through the face recognition camera or fingerprint scanning module. The system will generate a compressed summary vector locally and map it to a 64-bit hash string in a locally aware hashing manner and submit it to the on-chain comparison engine. The system extracts the biometric hash root value in the BioBinding field from the DID document bound to the user, and calculates the Hamming distance to determine whether the verification standard is met. If the user's behavior environment is stable, the device is consistent, and the DID confidence is normal, the system will quickly return the authorization result and the business window will release it directly; if there is remote access, frequent device switching or other risk signals, the system will initiate a multi-factor re-verification process, prompting the user to supplement other modalities or perform secondary verification, and automatically adjust the DID confidence based on the behavior map.

[0154] According to operational data statistics for six months after deployment, the system has demonstrated significant advantages in both verification accuracy and user access efficiency. Compared to the original system, the false rejection rate for biometrics has been reduced from 2.7% to 0.6%, and the multimodal verification pass rate has increased to 99.2%. In terms of identity authentication time, the average time it takes to enter the DID for the first time is about 14.8 seconds, and re-verification only takes 2.6 seconds to complete the judgment. More notably, the on-chain identity structure has good traceability and version control capabilities. On average, each user performs an identity update 1.7 times within six months, but there are no records of conflicts or identity reconstruction failures in the entire system's on-chain structure. In handling suspected abnormal access incidents, 52 high-risk logins were discovered through DID confidence sliding analysis, and 37 risk paths were successfully blocked. Four of the cases that assisted the public security in investigating involved identity fraud, all of which can be fully traced back through system logs.

[0155] This embodiment proves that the present invention can achieve the goals of data privacy protection, verification accuracy improvement and system performance optimization in high-concurrency and high-risk identity verification scenarios, and has strong practicality, good scalability and obvious commercial value.

[0156] The above description is only a preferred specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with the technical field, within the technical scope disclosed by the present invention, who makes equivalent replacements or changes based on the technical solution and inventive concept of the present invention, should be covered by the scope of protection of the present invention.

Claims

1. A rapid verification system integrating real-name DID and biometrics, characterized by: Includes the following modules: A biometric processing module, which is used to collect multimodal biometric data of users and generate nested biometric vectors through a variational nested encoder; A hash generation module, configured to perform a hash operation on the nested biometric feature vector and construct a Merkle tree to generate a biometric hash root value; Real-name verification module, used to verify the authenticity of the user's real-name information and generate real-name identity data; A DID building module is used to generate a DID document that complies with the W3C standard format and adds a BioBinding extension field to record the on-chain reference address of the biohash root value; a proof generation module for constructing a zero-knowledge proof based on the real-name identity data and the biometric hash root value, and generating a verification record; A blockchain writing module is used to write the DID document, BioBinding extension field, and verification record into the blockchain to form an on-chain real-name DID record; The verification module is used to collect user's single-modal lightweight biometric data during identity verification, generate a summary vector, and match it with the real-name DID record on the chain; The confidence control module is used to compare the matching results with the preset verification threshold. If the verification passes, the access is authorized; If the verification fails, the multi-factor re-verification mechanism will be triggered for a second verification and the DID confidence value will be calculated. If the second verification still fails, the identity will be frozen.

2. The rapid verification system integrating real-name DID and biometric features according to claim 1 is characterized in that: The modules are implemented as follows: S1. Collect the user's multimodal biometric data and input it into the variational nested encoder model to fuse and compress the mapping to generate a nested biometric vector; S2. Performing a hash operation on the nested biometric feature vector and constructing a Merkle tree to generate a biometric hash root value; S3. Obtain the real-name information provided by the user, verify the authenticity of the real-name information, and generate real-name identity data; S4. Generate a DID document in accordance with the W3C standard format based on the real-name identity data, and add a BioBinding extension field to record the on-chain reference address of the biohash root value; S5. Construct a zero-knowledge proof protocol input, use the zero-knowledge proof protocol to verify the binding relationship between the real-name identity data and the biometric hash root value, and generate a verification record; S6. Write the DID document, BioBinding extension field, and verification record into the blockchain to generate an immutable on-chain real-name DID record. S7. During identity verification, collect the user's single-modal lightweight biometric data, locally process the single-modal lightweight biometric data to generate a summary vector, and submit it to the on-chain comparison engine for matching with the on-chain real-name DID record; S8. Compare the matching result with the preset verification threshold. If the verification passes, the access is authorized. If the verification fails, the multi-factor re-verification mechanism will be triggered for a second verification and the DID confidence value will be calculated. If the second verification still fails, the identity will be frozen.

3. The rapid verification system integrating real-name DID and biometric features according to claim 2 is characterized in that: The multimodal biometric data includes facial images, fingerprint images and voiceprint signals.

4. The rapid verification system integrating real-name DID and biometric features according to claim 2 is characterized in that: Said S1 specifically includes: S11, the variational nested encoder model includes a modal feature extraction layer, a cross-modal fusion layer, a nested variational encoder layer and a compression mapping output layer; S12, inputting the multimodal biometric data into the modal feature extraction layer, performing convolutional coding, normalization, and activation processing on each modality, and extracting preliminary modal features; S13. Inputting the preliminary modal features into the cross-modal fusion layer, wherein the cross-modal fusion layer includes a multi-head attention mechanism and a gating mechanism for dynamically adjusting the weight relationship between the modalities to generate a modal fusion representation; S14. Input the modality fusion representation into the nested variational encoder layer, where the nested variational encoder layer consists of a main encoder and three sub-encoders, wherein the main encoder uses a multi-layer perceptron structure to model the modality fusion representation to extract the global potential mean and variance vector shared by all modalities; S15, the three sub-encoders correspond to facial images, fingerprint images, and voiceprint signals, respectively, use independent variational channels in the shared architecture to model local variation information of their respective modalities, and output modality-specific latent vectors; S16, fusing the global latent mean and variance vector output by the main encoder with the modal local latent vector output by each sub-encoder to generate a nested latent feature vector; S17: Input the nested latent feature vector into the compression mapping output layer, and obtain a structure-preserving nested biological feature vector through nonlinear dimensionality reduction mapping.

5. The rapid verification system integrating real-name DID and biometric features according to claim 2 is characterized in that: The S2 specifically includes: S21. Structurally divide the nested biometric feature vector into a number of sub-vector segments according to modality source and feature level, where each sub-vector segment corresponds to a feature set in a biometric modality or encoding level, for use in distributed hash calculations; S22. Perform three rounds of hashing operations on each subvector fragment using the SHA-3 hash algorithm. The first round performs hashing after vector normalization, the second round performs hashing after vector dimension rotation, and the third round performs scrambled hashing after introducing a mask parameter. Multiple hash subvalues are generated as leaf node values of the Merkle tree. S23. Construct the underlying structure of the Merkle tree. Use a sequential pairing method to concatenate two adjacent hash sub-values and then execute the SHA-3 hash algorithm again to generate the corresponding parent node hash value. Iterate upward layer by layer to generate intermediate nodes until the root node is generated. S24. During the construction of each layer of the Merkle tree, a modality identification tag is attached to each intermediate node. The modality identification tag represents the modality source combination contained in the intermediate node through a coding mapping method, which is used to trace the modality composition relationship; S25. A local update mechanism is introduced during the construction process. The input hash combination of each intermediate node is stored in a local buffer and an index table is constructed. When a subvector fragment changes, only the corresponding branch and upper-level node are updated to avoid repeated construction of the entire Merkle tree. S26. After completing the calculation of all intermediate nodes, the biometric hash root value is finally output as the unique identification hash root value of the nested biometric feature vector.

6. The rapid verification system integrating real-name DID and biometric features according to claim 2 is characterized in that: The S4 specifically includes: S41. Construct a basic identity information object based on the real-name identity data. The basic identity information object includes a user identifier, an issuing institution identifier, an issuance timestamp, an expiration timestamp, and an identity identification code, and is organized in a key-value pair format. S42. Generate a DID document according to the W3C Decentralized Identity Technical Specification. The DID document includes a context field, an id field, an authentication field, and a verificationMethod field. The context field specifies the specification version, and the id field corresponds to the user's DID identifier. S43. Set the elliptic curve-based public key identification data in the verificationMethod field, and set the signature method identifier for identity authentication in the authentication field; S44. Add an extended field BioBinding to the DID document. The BioBinding field includes four subfields: bioMerkleRoot, bindingTime, dataType, and storageReference, which correspond to the biometric hash root value, binding time, feature data type, and on-chain storage address information of the nested biometric feature vector, respectively. S45. Encapsulate the DID document containing the BioBinding extension field into JSON-LD format.

7. The rapid verification system integrating real-name DID and biometric features according to claim 2 is characterized in that: The S5 specifically includes: S51. Construct an input data set for the zero-knowledge proof protocol, and represent the real-name identity data and the biometric hash root value as the private input of the relationship to be proved; S52. Introduce zero-knowledge proof protocol verification, define constraint conditions based on the relationship to be proved, and express the constraint conditions in the form of a circuit, wherein the circuit form includes the constraint logic of the relationship between the real-name identity data and the biometric hash root value; S53. Generate a zero-knowledge proof public-private key pair based on the constraints of the circuit form, where the private key is used to generate the proof and the public key is used for on-chain verification; S54: Use the private key to perform zero-knowledge proof on the binding relationship between the real-name identity data and the biometric hash root value, and generate proof data, wherein the proof data includes a proof subject, a random challenge number, and a response factor; S55. Structured encoding of the certification data is performed, and digital signature is performed; S56. Combine the digitally signed proof data with the corresponding public key to form a verification record as a verification record of the binding relationship between the real-name identity data and the biometric hash root value. The verification record includes a signature summary and timestamp information.

8. The rapid verification system integrating real-name DID and biometric features according to claim 2 is characterized in that: When writing the DID document, BioBinding extension field and verification record to the blockchain, the pre-deployed smart contract registration interface is called, and the real-name DID record on the chain is used as the transaction payload and recorded in the block event log.

9. The rapid verification system integrating real-name DID and biometric features according to claim 2 is characterized in that: The S7 specifically includes: S71. During identity verification, collect single-modality lightweight biometric data of the user in real time through the terminal device, and perform scale normalization, noise filtering, and standardization on the single-modality lightweight biometric data to obtain a feature preprocessing vector. S72. A fixed random projection matrix and a bias vector are stored locally on the terminal device, wherein the random projection matrix is a predefined random value matrix, and the bias vector is a predefined random constant vector. A summary vector is generated by performing a nonlinear activation function ReLU mapping on the feature preprocessing vector. S73. For each element of the summary vector, perform binary quantization dimension by dimension using the sign function of the local-aware hash algorithm, convert each element into a corresponding binary sign bit, and form a local-aware hash string with 64 bits: Among them, h i Represents the value of the i-th bit in the local perception hash string, s i represents the i-th element in the summary vector; S74. Establish a communication connection with the blockchain network through the terminal device, encapsulate the local perception hash string as an on-chain transaction payload, and submit it to the on-chain comparison engine; S75. The on-chain comparison engine reads the on-chain real-name DID record, parses the DID document to verify the validity of the source, extracts the biometric hash root value in the BioBinding extension field, and verifies the signature summary and timestamp information in the record, and receives the local perception hash string and matches it with the biometric hash root value; S76. The on-chain comparison engine uses the Hamming distance function to calculate the similarity d between the local perception hash string and the biological hash root value. H : Among them, H bio,i Indicates the value of the i-th bit of the biometric hash root value, and d H Output as matching results.

10. The rapid verification system integrating real-name DID and biometric features according to claim 2 is characterized in that: The multi-factor re-verification mechanism specifically includes calculating the DID confidence value based on the user device type, historical login location, access frequency and verification behavior trajectory, and calculating the DID confidence change rate based on a sliding time window. When the confidence change rate exceeds the set confidence threshold, it is determined that the secondary verification still fails and the identity is frozen.

Images