Multimodal biometric fusion system based on one-time key protection

Abstract

The present invention relates to the field of key technology, and more specifically, to a multimodal biometric fusion system based on one-time key protection. It includes a revocable biometric conversion unit, a key guide fusion unit, a one-time protection unit, a key recovery unit, and a dual authentication unit. The present invention enriches the diversity of template fusion schemes on the one hand, and increases the difficulty for attackers to attack the system on the other hand. The generation of fusion templates is related to keys, and the keys are distributed and stored in different databases in a shared form. The keys can be restored losslessly only when all the shares are obtained, so as to generate hybrid templates. Both biometrics and keys are indispensable, and jointly ensure the security of the system and the privacy of users, and will not be forgotten and are not easy to be forged.

Description

Technical Field

[0001] The present invention relates to the field of key technology, and in particular to a multi-modal biometric fusion system based on one-time key protection. Background Art

[0002] With the rapid development of information networks, their influence on social life is becoming increasingly profound. From daily shopping and financial transactions to remote work and social media interactions, accurate and secure identity recognition technology has become the first line of defense to protect personal privacy and data security. Therefore, exploring and implementing efficient and reliable identity recognition methods has become an important issue and urgent need for current social development.

[0003] Among the existing user identification technologies, two commonly used technologies are: the traditional account plus key-based authentication method and the biometric-based authentication method. The traditional authentication method requires users to remember their usernames and keys. Although it provides users with a relatively simple and low-cost security guarantee, users often find it difficult to remember the key due to the complexity of key settings, or face the risk of being cracked by hackers due to insufficient key strength. These security risks not only threaten the security of users' personal information, but also hinder the further development of online identity authentication technology. Users still need a safe and reliable online identity authentication method.

[0004] In recent years, the fields of artificial intelligence, machine learning, computer vision, etc. have made great progress, and biometric technology has become the mainstream method of identity authentication. Biometrics uses machines to obtain unique behavioral or physiological characteristics of the human body to authenticate personal identity. These include gait, signature, fingerprint, iris, and face. Compared with traditional identity authentication methods based on accounts and keys, biometrics are completely bound to users and difficult to separate, so they will not be forgotten and are not easy to forge. However, the recognition accuracy of a single biometric recognition system is low and there are high application limitations. In order to solve this problem, research has gradually turned to multimodal biometric recognition technology that combines multiple biometrics.

[0005] Multimodal biometric authentication systems overcome common problems such as the security and non-universality of single systems. However, as an important task in multimodal biometrics, the fusion of biometrics still faces challenges. The data formats, dimensions, and distribution characteristics of different biometrics may vary greatly, and complex algorithms need to be designed to achieve effective data fusion. At the same time, this technology involves the collection and processing of multiple biometric information. Once this information is leaked or abused, it will pose a serious threat to personal privacy. In view of this, we propose a multimodal biometric fusion system based on one-time key protection. Summary of the invention

[0006] The purpose of the present invention is to provide a multimodal biometric fusion system based on one-time key protection to solve the problems raised in the above background technology.

[0007] To achieve the above-mentioned object, the present invention provides a multimodal biometric fusion system based on one-time key protection, comprising a revocable biometric conversion unit, a key boot fusion unit, a one-time protection unit, a key recovery unit and a dual authentication unit;

[0008] The revocable biometric conversion unit is used to sense the user's application for account registration signal during the registration phase, the administrator responds and requests to obtain the biometric feature image of the registered user, extract the multimodal feature vector through the feature extractor, and input the multimodal feature vector into the revocable biometric scheme for conversion in sequence to obtain the corresponding revocable template;

[0009] The key guidance fusion unit is used to randomly generate a key associated with the current user, and fuse each revocable template based on the value of each bit of the key to obtain an authentication fusion template;

[0010] The one-time protection unit is used to generate an OTP token by fusing the template and the key with the symmetric key assigned by a trusted third party, which is used as the verification information of the current registered user and is stored in the server;

[0011] The key recovery unit is used to obtain the biometric feature image of the verified user during the verification phase, process it through the feature extractor and encoder in sequence to obtain a binary vector, generate a revocable biometric feature again, recover the key from the share based on multi-party secure communication, and combine the recovered key with the revocable biometric feature template to generate a current hybrid template;

[0012] The dual authentication unit is used to input the current hybrid template into the server and match it with the authentication fusion template. If the verification matches, it outputs that the current user identity authentication is successful. If the verification does not match, it outputs that the current user identity authentication fails.

[0013] As a further improvement of the technical solution, the revocable biometric conversion unit includes a registration perception module, a biometric identification module and a feature extraction module;

[0014] The registration sensing module is used to receive an account registration request sent by a user to the management party, and respond to the registration request by keying out an account information filling page;

[0015] The biometric feature recognition module is used to control the camera of the management terminal device to turn on when sending a biometric feature request to the user, and to collect biometric feature images and transmit them to the server. The biometric feature images include face images, finger vein images, fingerprints, palm prints, irises and palm vein images;

[0016] The feature extraction module is used to drive the server to use a feature extractor to process the biometric image, extract a multimodal feature vector, and use a revocable biometric transformation algorithm to generate a set of random transformation parameters for each user, and use the transformation parameters to transform the multimodal feature vector to generate a revocable template.

[0017] As a further improvement of the present technical solution, the feature extraction module also includes a feature compatibility module, which is used to receive multiple biometric image features extracted by the feature extraction module and project biometric features of different modes into the same feature space without being limited by the size of the original biometric vector and the feature extractor.

[0018] As a further improvement of the technical solution, the key-guided fusion unit fuses the revocable templates based on the value of each bit of the key to obtain the authentication fusion template, including the following steps:

[0019] Given the length of the sliding window , the number of sliding windows , sliding step length , and the key length and the number of sliding windows than ,in, ;

[0020] Use sliding window to move each time For revocable templates Downsample and get the template:

[0021] ;

[0022] Use sliding window to move each time For revocable templates Downsample and get the template:

[0023] ;

[0024] Randomly generate a user-specific key , and based on the value of each bit of the key and Fusion is performed to obtain a fusion template , If the value of the bit is 0, then In As of , if If the value of is 1, then In As of .

[0025] As a further improvement of the technical solution, the one-time protection unit includes a splicing storage module, a symmetric key generation module and an update symmetric key module;

[0026] The splicing storage module is used to splice the mixed template and the key to generate a data block;

[0027] The symmetric key generation module is used to use the registration center RC as a trusted third party to generate a symmetric key using a hash function, and use an XOR operation to encrypt the combined data block to generate an OTP token;

[0028] The symmetric key updating module is used to transmit an update signal to the symmetric key generating module to update the symmetric key after each authentication is completed.

[0029] As a further improvement of the present technical solution, the one-time protection unit also includes a key distribution management module, which is used to use Shamir's secret sharing scheme to divide the key into multiple shared parts, where a threshold number of shared parts can restore the key, but less than the threshold number of shared parts cannot restore the key.

[0030] As a further improvement of the technical solution, the key recovery unit generates a current hybrid template, including the following steps:

[0031] The user sends an access request, so that the management end receives the access request, and again uses the biometric feature recognition module to collect multimodal biometric features, extracts the biometric feature vector of the current user, and generates a revocable biometric feature template according to the revocable conversion scheme through the feature extraction module;

[0032] The key is recovered from the multi-party secure communication sharing based on the key distribution management module, and the current hybrid template is generated again under the guidance of the recovered key through the key guidance fusion unit.

[0033] As a further improvement of the technical solution, the dual authentication unit matches the current mixed template input into the server with the authentication fusion template through a matcher, including the following gestures:

[0034] Posture 1: If multiple biometric features are matched successfully, the authentication of the biometric identification stage is output, and key authentication is performed. After the key authentication is passed, the dual authentication is output successfully;

[0035] Posture 2: After one of the biometric features of the current user fails to match, re-collect the biometric features for verification;

[0036] Posture 2: When all the user's biometric identification information fails to match, the current user's access request is rejected.

[0037] Compared with the prior art, the present invention has the following beneficial effects:

[0038] 1. In the multimodal biometric fusion system based on one-time key protection, by converting the original biometric into a revocable template, even if the attacker obtains the stored data, he cannot directly restore the user's original biometric, thereby preventing replay attacks and man-in-the-middle attacks. In addition, the use of revocable templates also overcomes the inherent irrevocability problem of biometrics, that is, once the biometric data is leaked, the user can "revoke" the old template and generate a new template without changing the biometric. The introduction of the key enriches the diversity of template fusion schemes on the one hand, and increases the difficulty of attackers attacking the system on the other hand. The generation of fusion templates is related to the key, and the key is distributed and stored in different databases in a shared form. The key can be restored losslessly only when all the shares are obtained, so as to generate hybrid templates. Both biometrics and keys are indispensable to jointly ensure the security of the system and user privacy.

[0039] 2. In the multimodal biometric fusion system based on one-time key protection, the fusion template and key are combined with the symmetric key assigned by the trusted third party through a one-time protection unit to generate an OTP token, which is used as the verification information of the current registered user and stored in the server, so as to realize further encryption operation of the hybrid template and key to prevent attackers from destroying the system security and leaking user privacy in this attack context, and it will not be forgotten and is not easy to be forged. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 It is the overall structural principle block diagram of the present invention;

[0041] Figure 2 is a flow chart of the registration phase of the present invention;

[0042] Figure 3 This is a schematic diagram of the key guidance fusion unit of the present invention;

[0043] Figure 4 It is a flow chart of the verification phase of the present invention.

[0044] The meaning of each number in the figure is:

[0045] 100. Revocable biometric conversion unit; 200. Key boot fusion unit; 300. One-time protection unit; 400. Key recovery unit; 500. Dual authentication unit. DETAILED DESCRIPTION

[0046] The following will be combined with the accompanying drawings in the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0047] See also Figure 1-Figure 4 As shown, this embodiment provides a multimodal biometric fusion system based on one-time key protection, including a revocable biometric conversion unit 100, a key guide fusion unit 200, a one-time protection unit 300, a key recovery unit 400 and a dual authentication unit 500;

[0048] The revocable biometric conversion unit 100 is used to sense the user's application for account registration signal during the registration phase. The administrator responds and requests to obtain the biometric feature image of the registered user, extracts the multimodal feature vector through the feature extractor, and inputs the multimodal feature vector into the revocable biometric scheme for conversion in sequence to obtain the corresponding revocable template. By converting the original biometric feature into a revocable template, even if the attacker obtains the stored data, the original biometric feature of the user cannot be directly restored, thereby preventing replay attacks and man-in-the-middle attacks. In addition, the use of revocable templates also overcomes the inherent irrevocability problem of biometrics, that is, once the biometric data is leaked, the user can "revoke" the old template and generate a new template without changing the biometric feature;

[0049] The key-guided fusion unit 200 is used to randomly generate a key associated with the current user, and fuse each revocable template based on the value of each bit of the key to obtain an authentication fusion template. The introduction of the key enriches the diversity of template fusion schemes on the one hand, and increases the difficulty for attackers to attack the system on the other hand. The generation of the fusion template is related to the key, and the key is distributed and stored in different databases in a shared form. The key can be restored losslessly only when all the shares are obtained, so as to generate a hybrid template. Both biometrics and keys are indispensable to ensure the security of the system and user privacy;

[0050] The one-time protection unit 300 is used to generate an OTP token by fusing the template and key with the symmetric key assigned by a trusted third party as the verification information of the current registered user and storing it in the server, so as to implement further encryption operations on the hybrid template and key. Since in the proposed multimodal biometric recognition system, malicious attackers may obtain the hybrid template or related data in the database through hacking techniques and then forge the hybrid template, and this possibility is increasing with the advancement of hacking techniques, it is necessary to further encrypt the key to generate the OTP token to prevent attackers from destroying system security and leaking user privacy under such attack background;

[0051] The key recovery unit 400 is used to obtain the biometric feature image of the verified user during the verification phase, and obtain a binary vector after processing it through the feature extractor and the encoder in sequence, and generate a revocable biometric feature again, which is conducive to the subsequent multi-modal biometric authentication through the revocable biometric feature, and synchronously recover the key from the share based on multi-party secure communication, and combine the recovered key with the revocable biometric feature template to generate the current mixed template, which is conducive to the subsequent two-factor authentication. On the basis of multi-modal biometric authentication, key authentication is performed. Multi-modal biometrics and passwords must be correct at the same time to authenticate the legitimate user. Confidentiality and biometrics complement each other, which improves the security and user privacy of MBAS.

[0052] The dual authentication unit 500 is used to input the current hybrid template into the server to match it with the authentication fusion template. If the verification matches, it outputs that the current user identity authentication is successful and meets the basic authentication. If the verification does not match, it outputs that the current user identity authentication fails. Example 1

[0053] The revocable biometric conversion unit 100 includes a registration sensing module, a biometric identification module, and a feature extraction module;

[0054] The registration sensing module is used to receive the account registration request sent by the user to the management party, and respond to the registration request, and key out the account information filling page to require the user to fill in the account information and necessary user information;

[0055] The biometric identification module is used to control the camera of the management terminal device to turn on when sending a biometric request to the user, and collect biometric images and transmit them to the server. The biometric identification feature images include face images, finger vein images, fingerprints, palm prints, irises and palm vein images;

[0056] The feature extraction module is used to drive the server to use the feature extractor to process the biometric image, extract the multimodal feature vector, and use the revocable biometric transformation algorithm to generate a set of random transformation parameters for each user, and use the transformation parameters to transform the multimodal feature vector to generate a revocable template. Specifically, the feature extractor can use mature feature extraction algorithms such as LBP, SIFT, and deep learning models to extract feature vectors, and the working principle of the revocable biometric transformation algorithm is as follows:

[0057] The random numbers generated by CSPRNG are highly random and unpredictable, ensuring the security of the transformation parameters. The BioHash algorithm is used to perform a hash operation on the feature vector and the random transformation parameters to generate a fixed-length binary vector. Different binary vectors can be generated by using different random transformation parameters. Even if an attacker obtains multiple revocable templates, the original feature vector cannot be derived.

[0058] Assume that identity recognition is performed based on two biometric features of the user. Let the two biometric features be , , using revocable biometric schemes and auxiliary data Transform the eigenvectors separately and , thus outputting two revocable templates and ,but:

[0059] ;

[0060] Among them, the revocable biometric scheme can be extended to biometric template protection schemes such as BioHashing, Bloom Filter, IoM Hashing and other revocable biometric schemes that may be proposed later. The use of revocable biometric schemes not only overcomes the inherent irrevocability of the original biometrics, but also enables biometrics of different modalities to be converted into feature vectors with the same value range and the same length.

[0061] Next, the feature extraction module also includes a feature compatibility module, which is used to receive multiple biometric image features extracted by the feature extraction module, and project biometric features of different modes into the same feature space without being restricted by the size of the original biometric vector and the feature extractor, thereby solving the problem of feature space incompatibility.

[0062] It is worth noting that the key guidance fusion unit 200 fuses the revocable templates based on the value of each bit of the key to obtain the authentication fusion template, including the following steps:

[0063] Given the length of the sliding window , the number of sliding windows , sliding step length , and the key length and the number of sliding windows than ,in, ;

[0064] Use sliding window to move each time For revocable templates Downsample and get the template:

[0065] ;

[0066] Use sliding window to move each time For revocable templates Downsample and get the template:

[0067] ;

[0068] Randomly generate a user-specific key , and based on the value of each bit of the key and Fusion is performed to obtain a fusion template , If the value of the bit is 0, then In As of , if If the value of is 1, then In As of ,in, , and They are all bit values. The randomly generated keys are highly random and unpredictable, ensuring the security of the keys. The length of the keys can be adjusted according to security requirements.

[0069] In order to further improve security, further encryption operations are performed on the mixed template Φ and the key secret, so that the one-time protection unit 300 includes a splicing storage module, a symmetric key generation module and an update symmetric key module;

[0070] The splicing storage module is used to splice the mixed template and the key to generate a data block;

[0071] The symmetric key generation module is used to use the registration center RC as a trusted third party to generate a symmetric key using a hash function to ensure the security and unpredictability of the symmetric key, and use an XOR operation to encrypt the combined data block to generate an OTP token to ensure that the combined data of the hybrid template and the key is effectively encrypted to prevent data leakage;

[0072] The symmetric key update module is used to transmit an update signal to the symmetric key generation module after each authentication is completed, update the symmetric key, ensure that each generated OTP token is unique, prevent replay attacks and man-in-the-middle attacks, and since the symmetric key is updated after each authentication is completed, attackers cannot use the old OTP token for replay attacks.

[0073] Furthermore, the one-time protection unit 300 further includes a key distribution management module, which is used to divide the key into multiple shared parts using Shamir's secret sharing scheme, wherein a threshold number of shared parts can restore the key, but less than the threshold number of shared parts cannot restore the key, and the multiple shared parts are stored in different databases to ensure that the leakage of a single database will not lead to the leakage of the key;

[0074] Among them, Shamir's secret sharing scheme is a commonly used secret sharing technology, which allows a key to be split into multiple shared values, among which any m shared values ​​can reconstruct the original key, m is a threshold, indicating how many shared values ​​are needed to reconstruct the key, and an m-1 degree polynomial can be selected to calculate m shared values, and each shared value is distributed to a participant, and the key is recovered from the share based on multi-party secure communication. The Lagrange interpolation method can be used to reconstruct the original key from any shared values, so that each participant can only obtain part of the key, but multiple participants can collaborate to reconstruct the original key. The technology not only improves the security of the key, but also prevents the key from being leaked by a single node or obtained by an attacker. In the key recovery phase, the Lagrange interpolation method can be used to reconstruct the original key from any shared values ​​to ensure the security and reliability of the system.

[0075] In summary, this embodiment is mainly disclosed in the registration stage, such as Figure 2 As shown, it can be summarized into the following steps:

[0076] First, in the revocable biometric conversion unit 100, the user sends an account registration request, the management end responds to the registration request, and types out an account information filling page to send to the user, so that the user fills in the account information. After sensing the account information filling, the management end sends a biometric request, so that the user collects a biometric image through the camera of the management end. After receiving all the data uploaded by the user, the management end can perform a preliminary manual review. After the biometric image standard is manually reviewed and the account information is filled in correctly, a pass signal is output to establish a registered account. Otherwise, an account registration failure is sent to the user. After the account is registered, the driver server uses a feature extractor to process the biometric image to extract a multimodal feature vector, and then uses a revocable biometric transformation algorithm to generate a set of random transformation parameters for each user, and uses the transformation parameters to transform the multimodal feature vector to generate a revocable template;

[0077] Next, the key-guided fusion unit fuses the revocable templates based on the value of each bit of the key to obtain the authentication fusion template, so that the one-time protection unit gradually differentiates the fusion template and the key with the symmetric key assigned by the trusted third party as the verification information of the current registered user and stores it in the server, so that the biometrics are completely bound to the user and are difficult to separate. Therefore, they will not be forgotten and are not easy to forge. They are easy to deploy on various platforms, providing users with a more convenient authentication experience. Example 2

[0078] In order to achieve verification, the user-side device interacts with the management-side device to authenticate the identity of the user who issues the access request, such as Figure 4 As shown, first, the key recovery unit 400 generates a current hybrid template, including the following steps:

[0079] The user sends an access request, so that the management end receives the access request, and again uses the biometric feature recognition module to collect multimodal biometric features, extracts the biometric feature vector of the current user, and generates a revocable biometric feature template according to the revocable conversion scheme through the feature extraction module;

[0080] The key is restored from the multi-party secure communication sharing based on the key distribution management module, and the current hybrid template is generated again under the guidance of the restored key through the key guidance fusion unit 200.

[0081] Furthermore, the dual authentication unit 500 matches the current mixed template input into the server with the authentication fusion template through the matcher, including the following gestures:

[0082] Posture 1. If multiple biometric features are matched successfully, the authentication of the biometric identification stage is output, and key authentication is performed. After the key authentication is passed, the double authentication is output successfully, realizing that multimodal biometrics and passwords must be correct at the same time to authenticate the legitimate user. Confidentiality and biometrics complement each other, which improves the security and user privacy of MBAS;

[0083] Posture 2: After one of the biometric features of the current user fails to match, re-collect the biometric features for verification to avoid failures caused by inaccurate collection and improve verification accuracy;

[0084] Posture 2: When all the user's biometric identification information fails to match, the current user's access request is rejected.

[0085] In summary, when performing the verification operation in the embodiment, the following detailed steps can be summarized, taking face and fingerprint as biometric features as an example:

[0086] The user enters the user ID and other information on the user-end device or the management-end device and sends an access request to the server-end device. The server-end device sends a command to control the camera of the management-end device to turn on, and the fingerprint collector to determine whether the current user's face is scanned and the fingerprint is collected. If the acquisition is successful, it enters the next step of the authentication phase; otherwise, it continues to scan and obtain biometric features;

[0087] The management device extracts the facial features and fingerprint features from the collected feature information and sends them to the server device. The server device generates a revocable template after processing. The server device also recovers the key from the share based on the secure communication of multiple parties and guides the biometric templates of fingerprint and face to generate a fusion template.

[0088] The server-side device determines whether the generated template matches the verification information of the corresponding user ID. If the match is successful, the server-side device returns access success to the management-side device and displays it on the display screen. If the match fails, the server-side device determines whether the number of times has reached the upper limit. If not, it sends a command to the management-side device to re-collect the face or other biometric features. The management-side device re-collects the biometric features. If the upper limit is reached, it sends a request to deny the user's access to the management-side device and displays it on the display screen;

[0089] Among them, the user-end device is any one of a mobile phone, tablet computer, laptop computer, desktop computer, or other special digital devices. The management-end device and the server-end device are independent and communicatively connected devices, similar to the structure of a bank's back-end server and front-end ATM.

[0090] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. The above embodiments and descriptions are only preferred examples of the present invention and are not intended to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A multimodal biometric fusion system based on one-time key protection, characterized by: It comprises a revocable biometric conversion unit (100), a key guidance fusion unit (200), a one-time protection unit (300), a key recovery unit (400) and a dual authentication unit (500); The revocable biometric conversion unit (100) is used to respond to and obtain the biometric feature image of the registered user during the registration phase, extract the multimodal feature vector, and input the multimodal feature vector into the revocable biometric scheme for conversion to obtain a corresponding revocable template; The key guidance fusion unit (200) is used to randomly generate a key associated with the current user, and fuse each revocable template based on the value of each bit of the key to obtain an authentication fusion template; The one-time protection unit (300) is used to generate an OTP token by fusing the template and the key with a symmetric key assigned by a trusted third party, which is used as the verification information of the current registered user and stored in the server; the key recovery unit (400) is used to obtain the biometric feature image of the verified user during the verification phase, regenerate a revocable biometric feature, and recover the key based on the secure communication of multiple parties, so that the recovered key is combined with the revocable biometric feature template to generate the current hybrid template; The dual authentication unit (500) is used to input the current mixed template into the server to match it with the authentication fusion template, and output the matching result.

2. The multimodal biometric fusion system based on one-time key protection according to claim 1 is characterized in that: The revocable biometric conversion unit (100) comprises a registration sensing module, a biometric identification module and a feature extraction module; The registration sensing module is used to receive an account registration request sent by a user to the management party, and respond to the registration request by keying out an account information filling page; The biometric feature recognition module is used to control the camera of the management terminal device to turn on when sending a biometric feature request to the user, and to collect biometric feature images and transmit them to the server. The biometric feature images include face images, finger vein images, fingerprints, palm prints, irises and palm vein images; The feature extraction module is used to drive the server to use a feature extractor to process the biometric image, extract a multimodal feature vector, and use a revocable biometric transformation algorithm to generate a set of random transformation parameters for each user, and use the transformation parameters to transform the multimodal feature vector to generate a revocable template.

3. The multimodal biometric fusion system based on one-time key protection according to claim 2 is characterized in that: The feature extraction module also includes a feature compatibility module, which is used to receive multiple biometric image features extracted by the feature extraction module and project biometric features of different modes into the same feature space without being limited by the size of the original biometric vector and the feature extractor.

4. The multimodal biometric fusion system based on one-time key protection according to claim 3 is characterized in that: The key guidance fusion unit (200) fuses the revocable templates based on the value of each bit of the key to obtain an authentication fusion template, comprising the following steps: Given the length of the sliding window , the number of sliding windows , sliding step length , and the key length and the number of sliding windows than ,in, ; Use sliding window to move each time For revocable templates Downsample and get the template: ； Use sliding window to move each time For revocable templates Downsample and get the template: ； Randomly generate a user-specific key , and based on the value of each bit of the key and Fusion is performed to obtain a fusion template , If the value of the bit is 0, then In As of , if If the value of is 1, then In As of .

5. The multimodal biometric fusion system based on one-time key protection according to claim 4 is characterized in that: The one-time protection unit (300) comprises a splicing storage module, a symmetric key generation module and a symmetric key update module; The splicing storage module is used to splice the mixed template and the key to generate a data block; The symmetric key generation module is used to use the registration center RC as a trusted third party to generate a symmetric key using a hash function, and use an XOR operation to encrypt the combined data block to generate an OTP token; The symmetric key updating module is used to transmit an update signal to the symmetric key generating module to update the symmetric key after each authentication is completed.

6. The multimodal biometric fusion system based on one-time key protection according to claim 5 is characterized in that: The one-time protection unit (300) also includes a key distribution management module, which is used to use Shamir's secret sharing scheme to divide the key into multiple shared parts, wherein a threshold number of shared parts can restore the key, but less than the threshold number of shared parts cannot restore the key.

7. The multimodal biometric fusion system based on one-time key protection according to claim 6 is characterized in that: The key recovery unit (400) generates a current hybrid template, comprising the following steps: The user sends an access request, so that the management end receives the access request, and again uses the biometric feature recognition module to collect multimodal biometric features, extracts the biometric feature vector of the current user, and generates a revocable biometric feature template according to the revocable conversion scheme through the feature extraction module; A key is restored from the multi-party secure communication sharing based on the key distribution management module, and a current hybrid template is generated again under the guidance of the restored key through a key guidance fusion unit (200).

8. The multimodal biometric fusion system based on one-time key protection according to claim 1, characterized in that: The dual authentication unit (500) matches the current mixed template input into the server with the authentication fusion template through a matcher, including the following gestures: Posture 1: If multiple biometric features are matched successfully, the authentication of the biometric feature recognition stage is output, and key authentication is performed. After the key authentication is passed, the dual authentication is output successfully; Posture 2: After one of the biometric features of the current user fails to match, re-collect the biometric features for verification; Posture 2: When all biometric identification information of the user fails to match, the access request of the current user is rejected.

Images