Biological characteristic authentication method and system for protecting user privacy
Through homomorphic encryption and privacy intersection technology, the problem of privacy leakage in biometric authentication system is solved, efficient and secure user privacy protection is achieved, computing costs are reduced and authentication accuracy is improved.
Patent Information
- Application Number
- CN202510631186.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-05-16
- Publication Date
- 2025-09-19
AI Technical Summary
The existing biometric authentication system cannot effectively protect user privacy, leading to privacy leakage problems.
By adopting homomorphic encryption and privacy intersection technology, through the collaborative work of the client, server and trusted third party, biometric encryption processing and similarity calculation are performed to protect the user's biometric and similarity data from being leaked.
It achieves all-round protection of user data privacy, reduces computing costs, improves authentication accuracy and efficiency, and ensures data security and privacy.
Smart Images
Figure CN120675690A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to a biometric authentication method and system for protecting user privacy, and belongs to the technical field of cryptography. Background Art
[0002] Homomorphic encryption is a public-key encryption system that allows addition and multiplication operations to be performed on data while it remains encrypted. Ciphertext input, intermediate calculation results, and the final result cannot be decrypted except by the holder of the private key, ensuring privacy and security.
[0003] Private intersection is a type of secure multi-party computation task that allows two or more parties to determine the intersection of sets without revealing any other data. A private intersection implementation based on homomorphic encryption can reduce the number of interactions required for private intersection, while its bundling capabilities make it suitable for small to medium-sized private intersection tasks.
[0004] Biometric authentication often uses methods such as manual identification, pattern recognition, machine learning, and deep learning to extract corresponding features from a user's biological data. Since similar biometric data inputs produce similar features, calculating feature similarity and setting a reasonable threshold can achieve highly accurate biometric authentication. However, conventional biometric authentication systems fail to protect biometric features, leading to privacy concerns. Summary of the Invention
[0005] The purpose of the present invention is to propose a biometric authentication method and system for protecting user privacy based on homomorphic encryption and privacy intersection. Homomorphic encryption is used to protect the characteristics based on the paradigm of biometric authentication, and it is applicable to privacy protection of authentication of any biometric feature.
[0006] To achieve the above object, the technical solution adopted by the present invention is as follows:
[0007] A biometric authentication method for protecting user privacy comprises the following steps:
[0008] 1) During the user registration phase, the client collects the user's biometrics and quantifies them. The client then uses the public key provided by a trusted third party to perform homomorphic encryption to obtain the encrypted features. The client also generates a user ID and sends both the user ID and the encrypted features to the server, which then stores them in the server's encrypted database.
[0009] 2) During the user authentication phase, the client collects the user's biometric features at the current moment and quantizes and encodes them according to the user's authentication request to generate encrypted features. At the same time, the server extracts the corresponding encrypted features from the secret database based on the user ID, performs homomorphic similarity calculation on the two encrypted features, and generates a homomorphic similarity ciphertext based on random numbers, which is sent to a trusted third party. The trusted third party decrypts the homomorphic similarity ciphertext and, in conjunction with the server, performs a homomorphic privacy intersection calculation on the decrypted result to generate a new ciphertext and send it to the client. The client decrypts the ciphertext and determines whether the authentication is passed based on the decryption result.
[0010] Furthermore, during the initialization phase, a trusted third party generates a private key, a public key, and a relinearization key for the secret database, of which the public key is sent to the client and the relinearization key is sent to the server. The client generates a private key and a public key for homomorphic privacy intersection and sends the public key to the server.
[0011] Furthermore, in the user registration stage, the step of quantizing and encoding the user's biometrics includes: multiplying the biometrics by an integer γ and rounding it to obtain the plaintext polynomial f0+f1x+…+f m-1 x m-1 .
[0012] Furthermore, in the user authentication stage, the step of quantizing and encoding the user's biometric characteristics at the current moment includes: multiplying the biometric characteristics by an integer γ and rounding it to obtain the plaintext polynomial g0-g m-1 x n-m+1 -g m-2 x n-m+2 -…-g2x n-2 -g1x n-1 .
[0013] Furthermore, the step of calculating homomorphic similarity includes: performing homomorphic multiplication on the two encrypted features c and c′ to obtain Then add the random number r to In the example, we get the homomorphic similarity ciphertext Where Δ is the encoding factor of homomorphic encryption.
[0014] Furthermore, the trusted third party uses the private key of the secret database to decrypt the homomorphic similarity ciphertext and obtains the polynomial decryption result sim+r.
[0015] Furthermore, the trusted third party and the server jointly perform homomorphic privacy intersection calculation on the encryption result, including the following steps:
[0016] The server adds the random number r to the threshold interval and scrambles the order, then encodes it into a polynomial in NTT form and encrypts it with the private intersection public key to obtain the ciphertext c t;
[0017] A trusted third party checks the ciphertext c t Subtract the polynomial sim+r to get c t -(sim+r,0); then additional N non-zero random numbers are randomly selected and encoded into a polynomial r′ in the form of NTT, and c t -(sim+r,0) multiplied to get the ciphertext (c t -(sim+r,0))·r′.
[0018] Furthermore, the step of judging whether the authentication is passed according to the decryption result includes: judging whether there is 0 in the decryption result, if there is 0, the authentication is passed, otherwise the authentication is failed.
[0019] A biometric authentication method for protecting user privacy, comprising:
[0020] The client is used to collect user biometrics, perform user registration and authentication, encrypt the collected biometrics and send them to the server, and make authentication decisions based on the received encryption results;
[0021] The server is used to store the secret database of user encrypted features, extract user encrypted features according to the query request during the user authentication phase, perform homomorphic similarity calculation with the currently collected encrypted features, and participate in homomorphic privacy intersection calculation with a trusted third party;
[0022] A trusted third party is used to decrypt the homomorphic similarity ciphertext and participate in the homomorphic privacy intersection calculation with the server.
[0023] The beneficial effects achieved by the present invention are as follows:
[0024] 1. This invention uses homomorphic similarity calculation and homomorphic privacy intersection to protect user biometrics, similarity, comparison thresholds, and other intermediate results, comprehensively protecting the data and privacy of all parties in the entire system.
[0025] 2. The homomorphic similarity calculation of the present invention only requires one homomorphic multiplication, and the homomorphic multiplication is optimized. The entire system has low computational cost, privacy and high precision.
[0026] 3. The homomorphic privacy intersection of the present invention only requires polynomial multiplication and addition to complete, and uses NTT to improve the efficiency of homomorphic similarity calculation, which is superior to other multi-round multi-party secure calculations. BRIEF DESCRIPTION OF THE DRAWINGS
[0027] Figure 1 This is a flow chart of biometric authentication for protecting user privacy in an embodiment of the present invention. DETAILED DESCRIPTION
[0028] In order to make the various technical features and advantages or technical effects of the above technical solutions of the present invention more obvious and easy to understand, they are described in detail below in conjunction with embodiments.
[0029] This embodiment provides a biometric authentication system for protecting user privacy, which adopts a three-party system including a client, a server and a trusted third party (CSP) that can communicate with each other. Figure 1 As shown, specifically including:
[0030] Client: Holds the private key for privacy exchange and decryption. It is the terminal device used by users for registration and authentication. It is responsible for collecting the user's biometrics and generating the user's encrypted features.
[0031] Server: holds a secret database, stores the user's secret features, uses homomorphic encryption algorithm to perform secret calculations on the similarity between features, and participates in homomorphic similarity calculations and homomorphic privacy intersection calculations.
[0032] Trusted third party: stores the decryption private key corresponding to the secret database and participates in the homomorphic privacy intersection calculation.
[0033] The server and a trusted third party jointly perform homomorphic privacy intersection calculations, using a homomorphic encryption algorithm to determine whether a value belongs to a specific threshold set in a confidential state.
[0034] The steps for biometric authentication based on the above system to protect user privacy are as follows:
[0035] 1. Initialization.
[0036] A trusted third party generates the private key, public key, and relinearization key corresponding to the secret database. The public key is sent to the client, and the relinearization key is sent to the server. These keys are used to compute similarity homomorphically. Furthermore, the client generates the private key and public key for homomorphic private intersection and sends the public key to the server.
[0037] 2. User feature collection.
[0038] The client collects user biometric data through the device and obtains user biometric features using feature extraction methods.
[0039] 3. User registration.
[0040] The client uses coefficient encoding to collect the user's biometric features, multiplies each feature by an integer γ and rounds it to quantize it, and constructs the plaintext polynomial f0+f1x+…+f m-1 x m-1 Use the public key provided by a trusted third party to encrypt the above plaintext polynomial to generate the user encryption feature Where n is a power of 2, and q is the ciphertext modulus. A user ID is also generated. The client sends this user ID along with the corresponding encrypted signature to the server, which then stores it in a secret database.
[0041] 4. Homomorphic calculation of similarity.
[0042] When the user initiates an authentication request, the client collects the user's biometric data again and uses the same feature extraction method to obtain the feature vector. The client encodes the newly collected features using coefficients to construct the plaintext polynomial g0-g m-1 x n -m+1 -g m-2 x n-m+2 -…-g2x n-2 -g1x n-1 The client encrypts the polynomial using the database public key to generate a new encryption signature And send it to the server.
[0043] The server queries the corresponding encrypted feature c from the secret database based on the ID provided by the user and performs homomorphic multiplication (See Figure 1 sim), where is the homomorphic multiplication symbol, the result modulo x n The constant term after +1 is the inner product of the two vectors. Then, in order to perform homomorphic privacy intersection calculation, the server takes a random number (random polynomial) r∈R q , calculate the homomorphic similarity ciphertext (See Figure 1 sim+r), where Δ is the encoding factor of homomorphic encryption, which is sent to a trusted third party.
[0044] 5. Homomorphic privacy intersection.
[0045] The server first adds the random number r to the threshold interval to obtain {θ+r,θ+r+1,…,θ+r+n}, scrambles the order, encodes it into a polynomial in NTT form, intersects the public key with privacy, and encrypts it into c t , sent to a trusted third party.
[0046] The trusted third party receives the homomorphic similarity ciphertext Finally, use the database private key to decrypt and get the polynomial sim+r, ensuring the data accuracy is around 0.01. Then add sim+r to the ciphertext c t Calculate the subtraction and get c t -(sim+r,0). Then additional N non-zero random numbers are randomly selected t is the plaintext modulus, which is encoded as a polynomial r' in NTT form. The plaintext and ciphertext multiplication is calculated to obtain the ciphertext (ct -(sim+r,0))·r′, sent to the client.
[0047] 6. User identity authentication.
[0048] After receiving the ciphertext returned by the server, the client uses the private key to decrypt it and determine whether the decrypted result contains a zero. If a zero appears in the decrypted result, it indicates that the similarity between the newly collected features and the registered features is within the legal threshold, and authentication succeeds. Otherwise, it indicates that the similarity is not within the legal threshold, and authentication fails.
[0049] Although the present invention has been disclosed as above by way of embodiments, they are not intended to limit the present invention. Any appropriate modification or equivalent substitution of the technical solution of the present invention by a person skilled in the art should be included in the protection scope of the present invention. The protection scope of the present invention shall be based on that defined in the claims.
Claims
1. A biometric authentication method for protecting user privacy, characterized in that: The following steps are involved: 1) During the user registration phase, the client collects the user's biometrics and quantifies them. The client then uses the public key provided by a trusted third party to perform homomorphic encryption to obtain the encrypted features. The client also generates a user ID and sends both the user ID and the encrypted features to the server, which then stores them in the server's encrypted database. 2) During the user authentication phase, the client collects the user's current biometric features based on the user's authentication request, quantizes and encodes them, and generates encrypted features. Simultaneously, the server extracts the corresponding encrypted features from the secret database based on the user ID, calculates the homomorphic similarity between the two encrypted features, and generates a homomorphic similarity ciphertext based on a random number, which is then sent to a trusted third party. The trusted third party decrypts the homomorphic similarity ciphertext and performs homomorphic privacy intersection calculation on the decryption result in conjunction with the server to generate a new ciphertext and send it to the client; the client decrypts the ciphertext and determines whether the authentication is passed based on the decryption result.
2. The method according to claim 1, wherein During the initialization phase, a trusted third party generates a private key, a public key, and a relinearization key for the secret database. The public key is sent to the client, and the relinearization key is sent to the server. The client generates a private key and a public key for homomorphic privacy intersection, and sends the public key to the server.
3. The method according to claim 1, wherein During the user registration phase, the steps for quantizing and encoding the user's biometrics include: multiplying the biometrics by an integer γ and rounding it to the integer, obtaining the plaintext polynomial f0+f1x+…+f m-1 x m-1 .
4. The method according to claim 1, wherein In the user authentication phase, the steps of quantizing and encoding the user's biometric features at the current moment include: multiplying the biometric features by an integer γ and rounding it to get the plaintext polynomial g0-g m- 1x n-m+1 -g m-2 x n-m+2 -…-g2x n-2 -g1x n-1 .
5. The method according to claim 1, wherein The steps of homomorphic similarity calculation include: performing homomorphic multiplication on two encrypted features c and c′ to obtain Then add the random number r to In the example, we get the homomorphic similarity ciphertext Where Δ is the encoding factor of homomorphic encryption.
6. The method according to claim 5, wherein The trusted third party uses the private key of the secret database to decrypt the homomorphic similarity ciphertext and obtains the polynomial decryption result sim+r.
7. The method according to claim 6, wherein The steps for the trusted third party to jointly perform homomorphic privacy intersection calculation on the encrypted results include: The server adds the random number r to the threshold interval and scrambles the order, then encodes it into a polynomial in NTT form and encrypts it with the private intersection public key to obtain the ciphertext c t ; A trusted third party checks the ciphertext c t Subtract the polynomial sim+r to get c t -(sim+r,0); then additional N non-zero random numbers are randomly selected and encoded into a polynomial r′ in the form of NTT, and c t -(sim+r,0) multiplied to get the ciphertext (c t -(sim+r,0))·r′.
8. The method according to claim 1, wherein The step of judging whether the authentication is passed according to the decryption result includes: judging whether there is 0 in the decryption result, if there is 0, the authentication is passed, otherwise the authentication is failed.
9. A biometric authentication method for protecting user privacy, configured to execute the method according to any one of claims 1 to 8, comprising: The client is used to collect user biometrics, perform user registration and authentication, encrypt the collected biometrics and send them to the server, and make authentication decisions based on the received encryption results; The server is used to store the secret database of user encrypted features, extract user encrypted features according to the query request during the user authentication phase, perform homomorphic similarity calculation with the currently collected encrypted features, and participate in homomorphic privacy intersection calculation with a trusted third party; A trusted third party is used to decrypt the homomorphic similarity ciphertext and participate in the homomorphic privacy intersection calculation with the server.