Virtual identity generation method based on meta universe

Abstract

The invention provides a virtual identity generation method based on meta universe, and aims to realize credible generation and virtual-real consistency verification of identities. According to the method, public and private keys of a user, iris features and semantic description information are collected, and multi-party credible binding is realized through an addition homomorphic threshold mechanism; a structure disturbance-table look-up mixed Hash mechanism is adopted to be fused with Griffin power mapping and Reinformed Concrete table look-up compression, and an anti-tamper abstract is generated for the user public key; physical identity information is generated in combination with the iris features and the proxy signature parameters; furthermore, through an identity-driven consistency twin network, behavior consistency verification and drive type discrimination between the virtual identity and the original identity are realized; the method is suitable for Web3.0, decentralized identity authentication and meta universe identity management, and has the advantages of high security, strong verifiability and cross-platform intercommunication capability.

Description

Technical Field

[0001] The present invention relates to the technical field of digital identity generation, and in particular to a method for generating a virtual identity based on a metaverse. Background Art

[0002] With the development of metaverse technology and the integration of related technologies such as virtual reality, augmented reality, and human-computer interaction, identity recognition and trusted authentication in virtual spaces have become increasingly core research and application focuses. Traditional identity authentication systems, which primarily rely on static credentials (such as usernames and passwords), single-factor biometrics (such as fingerprints, faces, and irises), or public-private key signature mechanisms, have been widely deployed in real-world scenarios. However, in highly open and anonymous digital environments such as the metaverse, they face a series of challenges such as identity forgery, key tampering, identity drift, and the lack of traceability of proxy identities. Existing technologies often generate virtual identities based on registration and binding based on user-defined digital information (such as nicknames and avatars). These systems lack deep binding mechanisms to the user's real physical identity or multi-factor characteristics, making them susceptible to impersonation or abuse. Furthermore, some virtual identity systems based on single-point central authentication lack effective support for cross-platform on-chain identity migration, proxy interaction, and behavioral auditability. Furthermore, conventional digital signature mechanisms struggle to simultaneously meet the conflicting requirements of "virtual expressiveness" and "real identity traceability" in metaverse scenarios. Summary of the Invention

[0003] The present invention provides a virtual identity generation method suitable for the metaverse environment, aiming to solve the problems existing in the existing virtual identity system, such as high risk of identity forgery, weak real identity binding ability, untraceable proxy identity and difficulty in cross-platform identity authentication; the method integrates multi-source information such as user public and private keys, biometrics (such as iris), semantic description, etc., and constructs an Avatar digital identity with verifiability, traceability and virtual-real consistency through a multi-party collaborative identity binding mechanism, a structural perturbation-lookup hybrid hash mechanism and an identity-driven consistency twin network; the method introduces an additive homomorphic threshold mechanism to achieve multi-party trusted binding of semantic features and user identities, thereby improving anti-forgery capabilities; by integrating Griffin power map perturbation and Reinforced The Concrete lookup table compression mechanism performs nonlinear structural perturbations and summary generation on user public keys to effectively prevent key replacement attacks; constructs a SharedCNN embedding network and a driven discriminant network, introduces a joint comparative loss function of class center alignment, discriminant boundaries, and information entropy constraints, and performs consistency verification and driven type reasoning on Avatar identities; this method can be widely used in scenarios such as Web3.0, decentralized identity authentication (DID), virtual reality interaction, and metaverse proxy identity management, breaking through the technical bottlenecks of existing virtual identities that are untraceable, unverifiable, and lack a secure binding mechanism.

[0004] The present invention provides a method for generating a virtual identity based on a metaverse, the method comprising the following steps:

[0005] Step S1: Collect user identification, user iris features, user virtual description, user private key and user public key, and submit them to the identity provider. The identity provider generates a metaverse identity token based on this information.

[0006] Step S2: performing identity binding on the user's virtual description and the user's public key through a multi-party collaborative virtual identity binding method to form virtual identity information, which includes a virtual identity hash value;

[0007] Step S3: Combining the user's iris features, the user's private key, and the virtual identity hash value, a proxy signature mechanism is used to generate signature parameters. Specifically, the virtual identity hash value is divided by the digest value of the user's iris features, and the ratio is subjected to the private key power operation of the user's private key to obtain the signature parameters. The user's iris features and the signature parameters are combined to form the physical identity information.

[0008] Step S4: Combining the Metaverse identity token, virtual identity information, and physical identity information to generate an Avatar digital identity as the Metaverse-virtual identity;

[0009] Step S5: Establish a shared feature extraction network, optimize the feature differentiation capability of the shared feature extraction network, construct an identity-driven consistency twin network, and authenticate and infer the Avatar digital identity through the identity-driven consistency twin network to ensure that the identity in the metaverse is consistent with reality and traceable; the identity-driven consistency twin network includes Shared CNN1, Shared CNN2 and the driving discrimination network.

[0010] Furthermore, step S2 specifically includes the following steps:

[0011] Step S21: extracting the description summary of the user's virtual description through a standard hash function to generate a semantic feature summary, which is used as the semantic representation of the binding message in the subsequent identity binding process;

[0012] Step S22: Define the set of participant public keys and threshold conditions, and select a set of participants that meet the threshold conditions from the set of participant public keys. Ensure the credibility of identity binding requires collaboration by multiple entities, improve the credibility of the binding structure, and prevent single-point forgery. For each participant in the set of participants, use its public key and semantic feature summary to perform chameleon hash binding, generate identity binding fragments, and form the basic data of identity mapping.

[0013] Step S23: Based on the basic identity mapping data, an additive homomorphic threshold mechanism is used to perform thresholded additive aggregation on all participants, generating a unified aggregated public key, an aggregated random factor, and a global hash value to construct an intermediate identity binding result. The obtained intermediate identity binding result implements a unique hash binding of the semantic feature summary under multi-party trusted verification, effectively enhancing the tamper resistance of the identity mapping and the traceability in the metaverse environment.

[0014] Step S24: Utilize the structural perturbation mechanism of the Griffin hash function and the Reinforced Concrete lookup table reorganization mechanism, integrate power mapping, MDS diffusion, and subdomain lookup table compression path, and construct a structural perturbation-lookup table hybrid hash mechanism. Use the structural perturbation-lookup table hybrid hash mechanism to perform summary calculation on the user public key to obtain a public key summary, which is used to prevent key replacement attacks; combine the public key summary with the global hash value and the aggregated random factor to construct a ternary binding structure, and then combine it with the semantic feature summary to construct a quaternary binding structure to obtain virtual identity information.

[0015] Furthermore, the process of authenticating and reasoning about the Avatar digital identity through the identity-driven consistency twin network specifically includes the following steps:

[0016] Step S51: Collect the original identity sample as a comparison benchmark, input the original identity sample and the Avatar digital identity into Shared CNN1 and Shared CNN2 for embedding mapping, and generate a feature vector pair;

[0017] Step S52: constructing a center alignment-entropy constraint joint contrast loss function through class center alignment, discrimination boundary and information entropy constraint, and calculating the intra-class consistency and inter-class separability index of the feature vector pair through the center alignment-entropy constraint joint contrast loss function to perform identity authentication;

[0018] Step S53: After identity authentication is confirmed, the micro-feature data of the Avatar digital identity is extracted and input into the drive discrimination network for inference and judgment to generate a judgment result; the judgment result includes human drive and AI agent drive; the micro-feature data includes iris texture data, facial expression dynamic change parameters, voice frequency domain fluctuation data and interaction rhythm stability data.

[0019] By adopting the above scheme, the beneficial effects achieved by the present invention are as follows:

[0020] The present invention establishes a collaborative binding relationship between the user's virtual description and the public keys of multiple parties by introducing an additive homomorphic threshold mechanism, realizing the identity mapping process of semantic feature summaries with the joint participation of multiple trusted entities; compared with traditional single-point signature or static binding methods, this mechanism improves the anti-counterfeiting ability of the user's virtual identity, ensuring that even if some participants are attacked, the overall binding result still has integrity and non-repudiation; this mechanism is particularly suitable for the distributed identity management needs in the metaverse, so that the generated Avatar digital identity has a multi-party collaborative verification basis, significantly improving the credibility and robustness of the virtual identity system in an open environment.

[0021] In the process of generating public key digests, the present invention integrates the Griffin structural perturbation mechanism and the ReinforcedConcrete lookup table reconstruction mechanism to construct a structural perturbation-lookup table hybrid hash mechanism adapted to zero-knowledge proof scenarios; this mechanism combines power mapping perturbation, MDS diffusion and subdomain-level lookup table mapping paths to effectively improve the structural distribution randomness and digest compression efficiency of user public keys in the digest processing process, and solves the problems of traditional hash algorithms such as insufficient structural perturbation, limited digest security and vulnerability to key replacement attacks; the digest generated by this mechanism has high nonlinearity, high collision resistance and low constraint complexity, providing an efficient and secure summary basis for the unique authentication of Avatar digital identity, and greatly enhancing the structural security and computational feasibility of the identity system.

[0022] In the identity authentication and drive reasoning stage, the present invention constructs an identity-driven consistency twin network, introduces a Shared CNN structure and a drive discrimination network, and combines the joint contrast loss function of class center alignment, discrimination boundary enhancement and information entropy constraint to achieve high-precision consistency verification and drive attribute reasoning judgment between the Avatar digital identity and the real identity; this mechanism not only solves the problem of lack of behavioral verification path for virtual identities in existing systems, but also improves the ability to discriminate complex identity behaviors (such as AI agent-driven and human-driven), and significantly enhances the system's interpretability, refined control capabilities and traceability; this design is particularly suitable for identity authentication needs in high-interaction, high-security scenarios such as Web3.0, metaverse social and AI agent interaction, ensuring that the virtual identity system achieves a robust connection between reality mapping and behavior control. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 This is a flow chart of a method for generating a virtual identity based on the metaverse proposed by the present invention;

[0024] Figure 2 This is a fusion diagram of constructing a quaternary binding structure proposed in Example 2. DETAILED DESCRIPTION

[0025] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments; based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0026] Example 1, according to Figure 1 The present invention provides a method for generating a virtual identity based on a metaverse, the method comprising the following steps:

[0027] Step S1: Collect identity: Collect user identity, user iris features, user virtual description, user private key and user public key, and submit them to the identity provider. The identity provider generates a metaverse identity token based on this information.

[0028] Step S2: Binding the virtual identity: Binding the user's virtual description and the user's public key through a multi-party collaborative virtual identity binding method to form virtual identity information, which includes a virtual identity hash value;

[0029] Step S3: Constructing a physical identity: Combining the user's iris features, the user's private key, and the virtual identity hash value, using a proxy signature mechanism, generates signature parameters. Specifically, the virtual identity hash value is divided by the digest value of the user's iris features, and the ratio is subjected to the private key power operation of the user's private key to obtain the signature parameters. The user's iris features and the signature parameters are combined to form the physical identity information.

[0030] Step S4: Generate Avatar Identity: Combine the Metaverse identity token, virtual identity information, and physical identity information to generate an Avatar digital identity as the Metaverse-virtual identity. The Avatar digital identity has both virtual behavior expression capabilities and traceable authentication capabilities, which are used to support subsequent login authentication, proxy authentication, and two-way mutual authentication application scenarios.

[0031] Step S5: Verification and reasoning: Establish a shared feature extraction network, optimize the feature differentiation capability of the shared feature extraction network, build an identity-driven consistency twin network, and authenticate and reason about the Avatar digital identity through the identity-driven consistency twin network to ensure that the identity in the metaverse is consistent with reality and traceable; the identity-driven consistency twin network includes Shared CNN1, Shared CNN2 and the driving discrimination network.

[0032] Example 2, according to Figure 2 This embodiment is based on the first embodiment. In this embodiment, step S2 specifically includes the following steps:

[0033] Step S21: extracting the description summary of the user's virtual description through a standard hash function to generate a semantic feature summary, which is used as the semantic representation of the binding message in the subsequent identity binding process;

[0034] Step S22: Define the set of participant public keys and threshold conditions, and select a set of participants that meet the threshold conditions from the set of participant public keys. Ensure the credibility of identity binding requires collaboration by multiple entities, improve the credibility of the binding structure, and prevent single-point forgery. For each participant in the set of participants, use its public key and semantic feature summary to perform chameleon hash binding, generate identity binding fragments, and form the basic data of identity mapping.

[0035] Step S23: Each participant includes a local hash fragment, a local public key, and a local random factor. Based on the basic data of identity mapping, an additive homomorphic threshold mechanism is used to perform threshold additive aggregation on the local hash fragments, local public keys, and local random factors of all participants, generate a unified aggregated public key, aggregated random factor, and global hash value, and construct an intermediate identity binding result. The obtained intermediate identity binding result realizes the unique hash binding of the semantic feature summary under multi-party trusted verification, effectively enhancing the tamper resistance of identity mapping and traceability in the metaverse environment. The formula used is as follows:

[0036] The formula used in the additive homomorphic threshold mechanism is as follows:

[0037]

[0038] Where T represents the set of participant indexes that meet the threshold, i represents the participant index, Represents the additive homomorphic aggregation operator, pk i represents the local public key of the i-th participant, and pk represents the public keys pk of all participants in the homomorphic addition aggregation T. i The obtained aggregate public key;

[0039]

[0040] Among them, r i represents the local binding random factor of the i-th participant, and r represents the random factor r of all participants. i Aggregate random factors obtained by homomorphic addition of ;

[0041]

[0042] Where h represents the local hash value h of all participants i Perform additive homomorphic aggregation to obtain the global hash value, h i represents the local chameleon hash fragment calculated by the i-th participant; CH pk(m, r) represents the aggregate chameleon hash operation on the semantic feature summary m using the aggregate public key pk and the aggregate random factor r;

[0043] Step S24: Construct a structural perturbation-table lookup hybrid hash mechanism, use the structural perturbation-table lookup hybrid hash mechanism to perform summary calculation on the user public key to obtain a public key summary, combine the public key summary with the global hash value and the aggregated random factor, and then combine it with the semantic feature summary to construct a four-element binding structure to obtain virtual identity information.

[0044] Example 3: This example is based on Example 2. In this example, step S24 specifically includes: utilizing the structural perturbation mechanism of the Griffin hash function and the Reinforced Concrete lookup table reorganization mechanism, integrating power mapping, MDS diffusion, and subdomain lookup table compression path, constructing a structural perturbation-lookup table hybrid hash mechanism, and using the structural perturbation-lookup table hybrid hash mechanism to perform summary calculation on the user public key to obtain a public key summary for preventing key substitution attacks; combining the public key summary with the global hash value and the aggregated random factor to construct a ternary binding structure, and then combining it with the semantic feature summary to construct a quaternary binding structure to obtain virtual identity information;

[0045] The specific steps of the structure perturbation-lookup hybrid hash mechanism to perform summary calculation on the user's public key are:

[0046] The vector of the user's public key is

[0047] Step 1: Power map perturbation layer (Griffin style):

[0048] Apply Griffin-type power mapping + polynomial perturbation function to the vector of the user's public key to obtain the vector after power mapping perturbation;

[0049] Step 2: MDS Diffusion

[0050] Use the MDS matrix and constant vector to perform MDS diffusion on the vector perturbed by the power map to obtain the diffused state vector;

[0051] Step 3: Domain lookup table compression path:

[0052] Decompose, replace and reorganize each element in the diffused state vector to obtain a compressed state vector;

[0053] Step 4: Concatenate compressed outputs:

[0054] Compress the compressed state vector through the sponge function to generate a public key digest;

[0055] Ternary binding structure:

[0056] VID core=(d,h,r);

[0057] Where d represents the public key summary, h represents the global hash value, r represents the aggregate random factor, and VID core represents a ternary binding structure;

[0058] Quaternary binding structure:

[0059] VID=(m,VID core );

[0060] Among them, m represents the semantic feature summary, VID represents the four-element binding structure;

[0061] The structural perturbation mechanism of the Griffin hash function uses a structural perturbation mechanism based on "position difference nonlinear mapping". Its core is to apply different power mapping transformations to different components of the input vector and introduce a perturbation strategy for state coupling to enhance nonlinear diffusion capabilities.

[0062] Reinforced Concrete lookup table reorganization mechanism: The Bars lookup table module is used to implement the structural compression transformation of state elements. The mechanism includes three steps: (1) decomposing the input domain elements into several subdomain chunks with smaller bit widths; (2) each chunk is mapped through a predefined nonlinear lookup table to complete the nonlinear perturbation; (3) finally, all the perturbed chunks are reorganized into complete finite field elements. Since the lookup table is essentially a constant time operation, this process has extremely low constraint complexity in the zero-knowledge proof (ZK) scenario and can be used for state compression or summary generation. It is an efficient structural encoding scheme.

[0063] Mechanism fusion: The structural perturbation mechanism of the Griffin hash function is integrated with the lookup and reorganization mechanism of Reinforced Concrete to construct a structural perturbation-lookup hybrid hash mechanism. In this mechanism, a Griffin-style power mapping perturbation and state coupling operation is first performed on the input user public key vector. That is, heterogeneous power function transformations are applied to the components separately, and a global structural deformation is introduced through a cross-perturbation function. Subsequently, the state vector after structural perturbation is input into the Bars module of Reinforced Concrete to complete the subdomain-level lookup mapping and reorganization compression, thereby efficiently generating a low-constraint summary value. Through this combination of pre-perturbation and post-compression, both structural randomness and constraint efficiency are improved, providing a robust, verifiable, and high-performance summary foundation for subsequent virtual identity binding.

[0064] Example 4. This example is based on Example 2. In this example, step S24 specifically includes: performing a conventional hash digest calculation on the user's public key to generate a corresponding public key digest value; combining the public key digest with the generated global hash value and the aggregated random factor to construct a ternary binding structure; on this basis, combining the semantic feature summary with the ternary binding structure to construct a quaternary binding structure, and finally generating virtual identity information.

[0065] Example 5. This example is based on Example 3. In this example, the process of authenticating and reasoning the Avatar digital identity through the identity-driven consistency twin network specifically includes the following steps:

[0066] Step S51: Collect the original identity sample as a comparison benchmark, input the original identity sample and the Avatar digital identity into Shared CNN1 and Shared CNN2 for embedding mapping, and generate a feature vector pair;

[0067] Step S52: Construct a center alignment-entropy constraint joint contrast loss function through class center alignment, discrimination boundary and information entropy constraint. Calculate the intra-class consistency and inter-class separability index of the feature vector pair through the center alignment-entropy constraint joint contrast loss function to perform identity authentication. The formula used is as follows:

[0068]

[0069] Among them, L info represents the information entropy contrast loss, η represents the information entropy loss weight coefficient, N represents the sample batch of Avatar digital identity, each sample corresponds to a feature vector pair; i represents the sample index, v i represents the embedding representation of the Avatar digital identity in the i-th feature vector pair, cy i represents the target identity category center corresponding to the i-th sample, corresponding to the original identity sample; K represents the total number of categories, j represents the category index, c j represents the j-th category center, τ represents the temperature coefficient, sim() represents the similarity function, exp() represents the exponential function,

[0070]

[0071] Where L represents the center alignment-entropy constraint joint contrast loss, L con represents the basic contrast loss term, λ represents the weight coefficient of the center boundary discrimination term, The embedding representation of the Avatar digital identity is the square of the Euclidean distance to the center of the target identity category, The embedding of the Avatar digital identity represents the square of the Euclidean distance to the center of the non-target identity category, Δ discriminant boundary interval, and []+ represents the ReLU truncation function;

[0072] Class center alignment:

[0073] Decision boundary:

[0074] Information entropy constraint:

[0075] Step S53: After identity authentication is confirmed, the micro-feature data of the Avatar digital identity is extracted and input into the drive discrimination network for inference and judgment to generate a judgment result; the judgment result includes human drive and AI agent drive; the micro-feature data includes iris texture data, facial expression dynamic change parameters, voice frequency domain fluctuation data and interaction rhythm stability data.

[0076] Example 6: This example is based on Example 3. In this example, the process of authenticating and reasoning the Avatar digital identity specifically includes the following steps:

[0077] Step S51: Collect the original identity sample as a comparison benchmark, input the original identity sample and the Avatar digital identity into Shared CNN1 and Shared CNN2 for embedding mapping, and generate a feature vector pair;

[0078] Step S52: Input the feature vector pair into the traditional contrast loss function to calculate the similarity score between the vectors, and judge the consistency between the Avatar digital identity and the original identity sample based on the similarity to complete the identity verification;

[0079] Step S53: After identity authentication is confirmed, the micro-feature data of the Avatar digital identity is extracted and input into the drive discrimination network for inference and judgment to generate a judgment result; the judgment result includes human drive and AI agent drive; the micro-feature data includes iris texture data, facial expression dynamic change parameters, voice frequency domain fluctuation data and interaction rhythm stability data.

[0080] Embodiment 7, this embodiment is based on embodiment 6. In this embodiment, step S4: combining the metaverse identity token, virtual identity information and physical identity information to generate an Avatar digital identity;

[0081] Avatar digital identity:

[0082]

[0083]

[0084] Step S5: Establish a shared feature extraction network, optimize the feature differentiation capability of the shared feature extraction network, build an identity-driven consistency twin network, and authenticate and infer the Avatar digital identity through the identity-driven consistency twin network;

[0085] Identity verification: The threshold is set to 0.1, and the center alignment-entropy constraint joint contrast loss = 0.04, which is less than the threshold, and the verification passes;

[0086] Reasoning and judgment:

[0087] The micro-feature data judgment is shown in Table 1:

[0088] Table 1

[0089]

[0090] Determined to be human-driven.

[0091] The present invention and its embodiments are described above. Such description is not restrictive. What is shown in the accompanying drawings is only one of the embodiments of the present invention, and the actual structure is not limited thereto. In short, if ordinary technicians in this field are inspired by it and do not depart from the purpose of the invention, they can creatively design structural methods and embodiments similar to the technical solution, which should all fall within the scope of protection of the present invention.

Claims

1. A method for generating a virtual identity based on the Metaverse, characterized by: The method comprises the following steps: Step S1: Collect user identification, user iris features, user virtual description, user private key and user public key to generate a Metaverse identity token; Step S2: performing identity binding on the user's virtual description and the user's public key through a multi-party collaborative virtual identity binding method to form virtual identity information, which includes a virtual identity hash value; Step S3: Combine the user's iris features, the user's private key, and the virtual identity hash value to generate signature parameters to form physical identity information; Step S4: Combining the Metaverse identity token, virtual identity information, and physical identity information to generate an Avatar digital identity as the Metaverse-virtual identity; Step S5: Build an identity-driven consistency twin network, and authenticate and infer the Avatar digital identity through the identity-driven consistency twin network.

2. The method for generating a virtual identity based on the metaverse according to claim 1, characterized in that: The identity-driven consistency twin network includes a feature mapping unit, a verification discriminant unit, and an inference discriminant unit. The feature mapping unit has built-in Shared CNN1 and Shared CNN2, and the inference discriminant unit has a built-in driving discriminant network.

3. The method for generating a virtual identity based on the metaverse according to claim 1, characterized in that: Step S2 specifically includes the following steps: Step S21: extracting the description summary of the user's virtual description and generating a semantic feature summary; Step S22: Define a set of participant public keys and a threshold condition, select a set of participants that meet the threshold condition from the set of participant public keys; perform chameleon hash binding on each participant in the set of participants using its public key and the semantic feature digest to form basic identity mapping data; Step S23: Based on the identity mapping basic data, an additive homomorphic threshold mechanism is used to perform threshold addition aggregation on all participants to generate an aggregated random factor and a global hash value; Step S24: Construct a structural perturbation-table lookup hybrid hash mechanism, use the structural perturbation-table lookup hybrid hash mechanism to perform summary calculation on the user public key to obtain a public key summary; combine the public key summary, global hash value, aggregated random factor and semantic feature summary to obtain virtual identity information.

4. The method for generating a virtual identity based on the metaverse according to claim 3, characterized in that: The structural perturbation-lookup hybrid hash mechanism is constructed using the structural perturbation mechanism of the Griffin hash function and the Reinforced Concrete lookup table reconstruction mechanism.

5. The method for generating a virtual identity based on the metaverse according to claim 2, characterized in that: The feature mapping unit collects original identity samples, inputs the original identity samples and Avatar digital identity into Shared CNN1 and Shared CNN2 for embedding mapping, and generates feature vector pairs.

6. The method for generating a virtual identity based on the Metaverse according to claim 5, characterized in that: The verification discriminant unit constructs a center alignment-entropy constraint joint contrast loss function, and calculates the index of the feature vector pair through the center alignment-entropy constraint joint contrast loss function to perform identity verification.

7. The method for generating a virtual identity based on the Metaverse according to claim 6, characterized in that: After the inference and discrimination unit confirms the identity authentication, it extracts the micro-feature data of the Avatar digital identity and inputs it into the driving discrimination network for inference and judgment to generate a judgment result.

8. The method for generating a virtual identity based on the Metaverse according to claim 7, characterized in that: Micro-feature data includes iris texture data, facial expression dynamic change parameters, voice frequency domain fluctuation data, and interaction rhythm stability data.

Images