A large-scale face biological information recognition encryption authentication method

Abstract

The application relates to the field of new-generation information technology and discloses a large-scale face biological information recognition encryption authentication method, which comprises the following steps: calling a two-party signature verification program in a networking server of a collection device to verify the authenticity of the collection device; calling an identification algorithm to verify the legality of the collection device; after confirming that the collection device is legal and authentic, establishing a true and multiple false transmission path; simultaneously performing homomorphic encryption on data content and generating a key; splitting the key into multiple key components and storing the multiple key components on nodes of a distributed network; performing logical operation on the data content without decryption, and completing the data original text by using a specific decryption method; verifying the legality of the collection device, encrypting the transmission path, and dynamically storing the key in a distributed manner, so that the difficulty of data theft and cracking is improved.

Description

Technical Field

[0001] This invention relates to the field of next-generation information technology, specifically to an encrypted authentication method for large-scale facial biometric information recognition. Background Technology

[0002] For facial biometric encryption, the prerequisite for facial biometric recognition is the completion of preliminary data collection. The collected information can be counterfeited and forged, making users' privacy and sensitive information vulnerable to attacks during facial recognition and transmission. Theft, alteration, or destruction of user data will cause significant losses to users. Due to the large scale and rapid growth of big data, its storage, encryption / decryption, signing, searching, and computation differ from traditional algorithms.

[0003] With the rapid development of the mobile internet in recent years, the ownership rate of mobile smart devices has increased significantly, and their usage scope has continued to expand. As a result, the importance of protecting user information security has also increased. However, due to the limited computing power of devices, the threat of key leakage has increased, and sensitive information stored in mobile devices has increasingly become a target of attacks. In the large-scale adoption of mobile smart devices for facial recognition, not only is the computing power of smart devices insufficient, but also the encryption methods are not rigorous enough. As a result, the collection, transmission, and storage of facial biometric information in large-scale facial recognition systems are all at risk of being attacked. Summary of the Invention

[0004] The purpose of this invention is to provide an encrypted authentication method for large-scale facial biometric information recognition. This method can effectively protect the security of facial biometric information while improving encryption efficiency and the convenience of key management.

[0005] A method for encrypted authentication of large-scale facial biometric information recognition is characterized by the following steps: With user authorization, a data acquisition device collects facial biometric information. When transmitting the facial biometric information, the device first calls a two-party signature verification program on a network server connected to the device to verify its authenticity. Simultaneously, a recognition algorithm is called from the server to verify the legitimacy of the device. If either authenticity or legitimacy fails verification, the device cannot transmit the facial biometric information. Once the device is confirmed to be legitimate and authentic, a one-true-multiple-false transmission path is established between the device and the receiving server. The facial biometric information is encrypted using a homomorphic encryption algorithm to form encrypted facial biometric information that can be processed in the encrypted state, and stored at the receiving end. A key is generated from the encrypted facial biometric information, which is then split into multiple key components and stored on nodes of a distributed network. When further processing of the facial biometric information is required, the encrypted facial biometric information is sent to the processing end.

[0006] The principle and advantages of this scheme are as follows: When using commonly used public data collection devices, such as mobile phones, network cameras, and cameras, to acquire facial biometric information on a large scale, the authenticity of the collection devices is first verified to ensure they are genuine. The legality of the data collected from these devices is also verified. During data transmission, the data transmission path is encrypted to prevent data transfer through physical interception. Furthermore, the key is split into multiple key components and stored in a distributed management system on nodes of a distributed network to prevent information loss and leakage. Homomorphic encryption allows data to be processed even in an unencrypted state, ensuring the security of the entire facial biometric information chain.

[0007] Preferably, the method for establishing a one-true-multiple-false transmission path includes simultaneously establishing multiple transmission paths in the acquisition device. Only one of these paths contains the correct data information, while the other paths contain ordinary encrypted fake data. The fake data extracts non-core portions of the real data. Both the correct and fake data embed a tracking program for the device's physical address. When the fake file is intercepted, the tracking program records the network address of the intercepted file and sends it to the monitoring server. The monitoring server then provides error text information for the intercepted address and sends an alert to the administrator. This ensures that data is not obtained through the transmission paths, delays data leakage through fake files, and allows the tracking program to trace the whereabouts of the stolen data.

[0008] Preferably, before the data acquisition device transmits data to the outside world, it also includes encryption authentication of the facial biometric information acquired by the acquisition device. By installing an acquisition client in the acquisition device, the acquisition device opens the client when acquiring facial biometric information. The client calls the two-party signature verification program in the network server to verify the authenticity of the acquisition device. At the same time, it calls the recognition algorithm from the server to verify the legality of the acquisition device. If either authenticity or legality fails the verification, the client cannot acquire facial biometric information.

[0009] Preferably, the key splitting and recovery process satisfies the following steps: First, a polynomial is randomly constructed for the user: ,in For user key, It is a prime number. , Then take A number of distinct random numbers: ,get indivual The key components, P, are then distributed among the n key recovery participants via a secure, encrypted SSL transmission channel. Finally, when key recovery is needed, the n key recovery participants send their respective key components... Substitute into the formula:

[0010] ,when hour:

[0011] Key .

[0012] The keys are split and stored separately to ensure that they are not easily cracked, thereby ensuring the security of facial biometric information.

[0013] Homomorphic encryption algorithms use an asymmetric method to generate public-private pairs. Security is ensured by using random variations in the encryption of the unchanging plaintext each time, ultimately achieving homomorphic properties. Homomorphic properties satisfy the following logical relationship:

[0014] ,in There are two plaintexts. For encryption operations, This is for decryption operations.

[0015] Preferably, in the process of recovering the key from facial biometric information, a key recovery based on (3, N) threshold sharing is adopted. Lagrange interpolation is used to achieve dynamic distribution and secure recovery of key components, where 3 represents the number of nodes available for key recovery and N represents the total number of nodes storing key components. .

[0016] This key-sharing method not only avoids the leakage of the complete key, but also enables the recovery of the key in the event of partial key loss.

[0017] Preferably, during the transmission of facial biometric information, a random factor is generated by the random number generator built into the acquisition device, and the random factor is encrypted together with the device information of the acquisition device using the SM2 algorithm to ensure the security of facial recognition information data during transmission.

[0018] Finally, the acquisition device can be a mobile phone, a PC, or an all-in-one machine with image acquisition capabilities, in order to expand the scope of application of this solution. Attached Figure Description

[0019] Figure 1 A flowchart for encrypted transmission of facial recognition data and keys;

[0020] Figure 2 A flowchart for key segmentation and storage;

[0021] Figure 3 A flowchart for signature authentication;

[0022] Figure 4 A flowchart for recovering a forgotten key;

[0023] Figure 5 A system functional architecture diagram for running the method of the present invention. Detailed Implementation

[0024] The present invention will be further described in detail below through specific embodiments, but the embodiments are not limited thereto. Unless otherwise specified, the technical means used in the following embodiments are conventional means well known to those skilled in the art.

[0025] Example 1

[0026] like Figure 1 As shown, under user authorization, when the acquisition device transmits facial biometric information, it first calls a two-party signature verification program in the network server connected to the acquisition device to verify the authenticity of the acquisition device; simultaneously, it calls a recognition algorithm from within the server to verify the legitimacy of the acquisition device. If either authenticity or legitimacy fails verification, the acquisition device cannot transmit facial biometric information. Once the acquisition device is confirmed to be legitimate and authentic, a one-true-multiple-false transmission path is established between the acquisition device and the receiving server. The facial biometric information is encrypted using a homomorphic encryption algorithm to form ciphertext that can be processed in the encrypted state, which is stored at the receiving end. A key is generated from the ciphertext of the facial biometric information, the key is split into multiple key components, and the multiple key components are stored on nodes of a distributed network. When further processing of the facial biometric information is required, the homomorphically encrypted facial biometric information is sent to the processing end.

[0027] The identification algorithm refers to a verification algorithm based on the hardware characteristics of the device, including the device's MAC address, factory serial number, encryption chip identifier, and a whitelist of devices pre-stored on the server. Its core function is to determine whether the device has the authority to collect and transmit facial biometric information by verifying whether the unique hardware identifier of the collection device has been legally registered on the server. The algorithm is implemented as follows: after the server receives the hardware identifier information sent by the collection device, it converts it into a fixed-length feature value through a hash operation and performs a consistency check with the feature values ​​stored in the whitelist. If the check passes, the device is deemed legitimate; otherwise, it is deemed illegitimate.

[0028] The method for establishing a one-true-multiple-false transmission path involves simultaneously establishing multiple transmission paths in the acquisition device. Only one of these transmission paths contains the correct data information, while the other transmission paths contain ordinary encrypted fake data. The fake data takes the non-core part of the real data and modifies the packaged data. Both the correct data and the fake data embed a tracking program for the device's physical address. When the fake file is intercepted, the tracking program records the network address of the intercepted file and sends it to the monitoring server. The monitoring server then provides error text information for the intercepted address and sends an alert to the administrator.

[0029] File spoofing tools can be used to modify the non-core parts of a real file and disguise it as another type of file, such as disguising a document as an image or audio file. This effectively prevents the file format from being easily detected and increases the credibility of the disguise.

[0030] To encrypt the disguised file, a symmetric encryption algorithm such as AES can be used to ensure that even if the disguised file is intercepted, its contents cannot be read without the correct key.

[0031] Multiple transmission paths are established, and network virtualization technology or proxy servers are used to create these fake transmission paths, making it difficult for attackers to determine which one is the real transmission path.

[0032] The tracking program automatically activates when the spoofed file is intercepted and opened. It utilizes unique identifiers such as the device's MAC address as tracking information. The tracking program is embedded within the spoofed file, ensuring it is triggered when the file is accessed. When the interceptor opens the spoofed file, the tracking program automatically activates, records the interceptor's network address and other information, and sends this information to the monitoring server.

[0033] The system uses specific network address access, file download, or opening as alert trigger conditions. Once an alert condition is triggered, the tracking program immediately sends an alert message to administrators, including detailed information such as the interceptor's network address and the interception time, so that administrators can take timely countermeasures.

[0034] Example 2

[0035] like Figure 1 and Figure 2 As shown, the data terminal requests authentication to verify the legitimacy of the device and generates a key pair through homomorphic encryption. The key pair is then encrypted using a cryptographic machine. The encrypted key is managed by first splitting the key into multiple key components and storing each component on a node in a distributed network. The key is transmitted via the SSL protocol. At the receiving end, decryption is performed through interactive computation to generate a public key.

[0036] In the key splitting method and the secure multi-party key management method, a threshold key sharing strategy based on Lagrange interpolation is used to recover the keys in the event of partial key loss.

[0037] Key component generation:

[0038] 1) User-generated random polynomials: ,in For user key, It is a prime number. , ;

[0039] 2): Take A number of distinct random numbers: ,get indivual A point is a key component.

[0040] Key component distribution method: via secure encryption SSL The transmission channel will n Key components P Send to separately n Key recovery participants.

[0041] The key recovery process is as follows: n Each key recovery participant will share their respective key components. Substitute into the formula:

[0042] ,when hour:

[0043] Key .

[0044] Homomorphic encryption algorithms generate public-private key pairs using an asymmetric method. Their core principle is to achieve homomorphic properties on top of an asymmetric encryption architecture, meaning that:

[0045] ,in There are two plaintexts. For encryption operations, This is for decryption operations.

[0046] like Figure 3As shown, homomorphically encrypted data can be processed by the server without decryption. The server obtains the ciphertext using a signature authentication method. This method begins with the server initiating a signature request, calculating the private key to obtain a matching signature value, thus verifying the signature's validity. The server then obtains the ciphertext of the facial biometric information from the server using the authorized signature. The server performs calculations on the ciphertext without decryption and uploads the calculated ciphertext to the server. The requesting party then retrieves the calculated ciphertext from the server and decrypts it to obtain the processed plaintext of the facial biometric information.

[0047] When generating an SM2 asymmetric key, two parties independently generate a private key share (or private key fragment, key component). They communicate and transmit auxiliary computational data, which is then combined by one party to generate the SM2 public key. As long as the two parties do not collude, it is impossible to recover the complete SM2 private key. Even if an attacker can only compromise one of the parties, it is also impossible for the attacker to recover the complete SM2 private key.

[0048] When it is necessary to sign the transmitted data, the two parties use their respective signature private key fragments to calculate and generate signature fragments. Then, the two parties exchange and transmit the signature fragments and other auxiliary calculation data, and one party performs the combined calculation on the received data.

[0049] When it is necessary to decrypt the ciphertext, the two parties use their respective decryption private key fragments to calculate and generate plaintext fragments. Then, the two parties exchange and transmit the plaintext fragments and other auxiliary calculation data. One party then merges and calculates the received data to generate the decrypted plaintext.

[0050] like Figure 4 As shown, if the key is forgotten, the way to recover the key is as follows: Initiate a request to forget the PIN code on the data end. After authenticating the user, the system obtains the node information where the password component is stored. Through the correlation between the key components, other key components are found. The key is obtained by reconstructing a polynomial based on multiple key components. A message to modify the PIN password is sent to the correctly authenticated user, and the PIN password is rebound to the key component.

[0051] Example 3:

[0052] like Figure 5 As shown, the browser installed on the data collection device achieves cross-browser compatibility through React and Vue, allowing stable and efficient access to the system without restricting browsing by any browser. It also achieves front-end and back-end separation, making the code architecture more independent.

[0053] The business layer includes external API interfaces, which facilitates quick and easy integration by third-party systems, as well as overall system data analysis and system performance analysis.

[0054] The key components are transmitted via HTTPS and SSL encrypted channels to ensure security during transmission. A gateway is used to unify authentication and authorization functions and implement protection policies. High availability is achieved using an Nginx reverse proxy.

[0055] The service layer uses the cross-platform language Java and employs Spring Boot to reduce deployment costs. MyBatis is used for communication with the database layer, and RocketMQ message queues are used to decouple some business logic and achieve high availability.

[0056] The data layer primarily uses the relational database MySQL and the non-relational databases Redis and MongoDB. It implements database read / write separation, data caching, cache expiration control, and transaction support.

[0057] The basic layer mainly includes systems such as hosts, networks, and storage, which support encryption and decryption, signature verification and calculation related to digital certificates.

[0058] The above descriptions are merely embodiments of the present invention, and common knowledge such as specific technical solutions and / or characteristics are not described in detail here. It should be noted that those skilled in the art can make various modifications and improvements without departing from the technical solutions of the present invention, and these should also be considered within the scope of protection of the present invention. These modifications and improvements will not affect the effectiveness of the implementation of the present invention or the practicality of the patent. The scope of protection claimed in this application should be determined by the content of its claims, and the specific embodiments described in the specification can be used to interpret the content of the claims.

Claims

1. A method for encrypted authentication of large-scale facial biometric information recognition, characterized in that, The process includes the following steps: With user authorization and consent, the data acquisition device collects facial biometric information. When the data acquisition device transmits the facial biometric information outward, it first calls a two-party signature verification program on the network server connected to the data acquisition device to verify the authenticity of the data acquisition device. At the same time, it calls a recognition algorithm from within the server to verify the legitimacy of the data acquisition device. If either authenticity or legitimacy fails verification, the data acquisition device cannot transmit the facial biometric information outward. Once the data acquisition device is confirmed to be legitimate and authentic, a one-true-multiple-false transmission path is established between the data acquisition device and the receiving server. The facial biometric information is then encrypted using a homomorphic encryption algorithm to form ciphertext of the facial biometric information that can be processed in the encrypted state, and stored at the receiving end. A key is generated from the encrypted facial biometric information. The key is then split into multiple key components, which are stored on nodes of a distributed network. When further processing of the facial biometric information is required, the encrypted facial biometric information is sent to the processing end. The method for establishing a one-true-multiple-false transmission path includes simultaneously establishing multiple transmission paths in the acquisition device. Only one of the multiple transmission paths contains the correct data information, while the other transmission paths contain ordinary encrypted fake data. The fake data is taken from the non-core part of the real data. Both the correct data and the fake data embed a tracking program for the device's physical address. When the fake file is intercepted, the tracking program records the network address of the intercepted file and sends it to the monitoring server. The monitoring server then provides error text information for the intercepted address and sends an alarm message to the administrator.

2. The encryption authentication method for large-scale facial biometric information recognition according to claim 1, characterized in that, Before the data acquisition device transmits data to the outside world, it also includes encryption authentication of the facial biometric information being collected. By installing a client in the acquisition device, the acquisition device opens the client when collecting facial biometric information. With the user's authorization and consent, the client calls the two-party signature verification program in the network server to verify the authenticity of the acquisition device. At the same time, it calls the recognition algorithm from the server to verify the legality of the acquisition device. If either authenticity or legality fails to pass the verification, the client cannot collect facial biometric information.

3. The encryption authentication method for large-scale facial biometric information recognition according to claim 1, characterized in that, The key splitting and recovery process follows these steps: First, a polynomial is randomly constructed based on the required number of key components. ,in For user key, It is a prime number. , Then take A number of distinct random numbers: ,get indivual The key components are then distributed among n key recovery participants via a secure, encrypted SSL transmission channel. When key recovery is needed, each of the n key recovery participants sends its respective key component. The key recovery operation logic is obtained by inputting the key into the key recovery operation logic, which is as follows: ,when hour: Key .

4. The encryption authentication method for large-scale facial biometric information recognition according to claim 3, characterized in that: In the process of recovering the key of facial biometric information, the key is recovered using a (3,N) threshold sharing method. The number of nodes that can be used to recover the key is 3, and the total number of nodes storing the key components is N, where N≥3. The key components are dynamically distributed and securely recovered using the Lagrange interpolation method.

5. The encryption authentication method for large-scale facial biometric identification according to claim 1 or 2, characterized in that, The two-party signature verification includes the following steps: the data sender and the data receiver each independently generate a private key share; the two parties communicate interactively, with the face biometric information receiver sending auxiliary data for signature verification to the face biometric information sender; the face biometric information sender then merges the auxiliary data to generate an SM2 public key, which is used to verify the authenticity of the data sender.

6. The encryption authentication method for large-scale facial biometric information recognition according to claim 1, characterized in that, The acquisition device can be a mobile phone, a PC, or an all-in-one machine with image acquisition capabilities.

Images