A blockchain-based facial recognition security system for campuses

Abstract

This application relates to the field of artificial intelligence technology and provides a blockchain-based campus facial recognition security system, comprising: a data processor for acquiring encrypted facial features of all employees of a corresponding enterprise and uploading the encrypted data to the enterprise's blockchain; a blockchain module for synchronizing the encrypted data in each enterprise blockchain to other enterprise blockchains through a consensus mechanism, and synchronizing the encrypted data in each enterprise blockchain to the campus blockchain through a smart contract; the blockchain module is also used to compare the hash value of the campus blockchain with the hash value of each enterprise blockchain through a smart contract. If the hash value of the campus blockchain is inconsistent with the hash value of any enterprise blockchain, a main chain is determined as the data synchronization standard, and the encrypted data stored in the main chain is synchronized to other chains. This application can prevent the access control system from being paralyzed when successfully attacked, and effectively protect the personal data security of employees.

Description

Technical Field

[0001] This application belongs to the field of artificial intelligence technology, and in particular relates to a blockchain-based campus facial recognition security system. Background Technology

[0002] In the digital age, cyberattacks have become a global problem, with their frequency and complexity constantly increasing. They not only threaten enterprise security management but can also lead to significant business losses and even hefty penalties in severe cases. In recent years, with the rapid development of internet technology, cyberattack methods have also evolved. These attacks not only damage data integrity and availability but may also lead to the leakage of sensitive information. Therefore, developing effective defense mechanisms to protect the network environment has become a crucial issue in the field of information technology. Simultaneously, with the popularization of laws such as the Cybersecurity Law, the Data Security Law, and the Personal Information Protection Law, employees' awareness of personal information protection has increased. Against the backdrop of frequent data breaches, employees are paying increasing attention to the security of their personal information. Enterprises bear a significant responsibility in maintaining employee information security while also facing enormous challenges.

[0003] Currently, facial recognition technology is widely used in various settings, from social media to security monitoring systems; it has become ubiquitous. As sensitive personal information, the leakage of facial data can lead to serious consequences such as identity theft and financial losses. In large industrial parks that use facial recognition access control, there are often many independent companies. To protect their personal information and prevent privacy breaches, employees may prefer to have their facial data registered with their employer rather than directly with park management. While individual companies typically have internal security departments responsible for the security of their employees' personal information, large industrial parks, acting as property management and security services providers, tend to have weaker cybersecurity and data security protections. Compared to companies within parks with independent security departments, attackers may find it easier to target the park. If an attacker successfully breaches the park's access control system, not only will a large amount of data be lost, causing the system to crash, but it could also lead to the leakage of substantial employee data, resulting in severe consequences.

[0004] In summary, the park's access control system currently suffers from issues such as system paralysis and personal data leakage when successfully attacked. Summary of the Invention

[0005] This application provides a blockchain-based facial recognition security system for campuses, which can solve the problems of access control system paralysis and personal data leakage when the campus access control system is successfully attacked.

[0006] This application provides a blockchain-based facial recognition security system for campuses, including:

[0007] Multiple data processors are associated with multiple companies within the park. The data processors are used to obtain encrypted facial features of all employees of the company corresponding to the data processor and upload the encrypted data to the corresponding company's blockchain.

[0008] The blockchain module is used to synchronize the encrypted data in each enterprise blockchain to other enterprise blockchains through a consensus mechanism, and to synchronize the encrypted data in each enterprise blockchain to the park's blockchain through smart contracts, so that the encrypted data of the facial features of all employees of all enterprises are stored in the enterprise blockchains and the park blockchain; all enterprise blockchains and the park blockchain are in the same consortium blockchain network.

[0009] The blockchain module is also used to compare the hash value of the park blockchain with the hash value of each enterprise blockchain through smart contracts. If the hash value of the park blockchain is inconsistent with the hash value of any enterprise blockchain, the main chain is determined from the park blockchain and all enterprise blockchains to serve as the data synchronization standard, and the ciphertext stored in the main chain is synchronized to the other chains in the park blockchain and all enterprise blockchains except the main chain.

[0010] Optionally, the data processor is specifically used to: obtain the facial features of all employees of the enterprise corresponding to the data processor, and encrypt each facial feature using a homomorphic encryption public key to obtain the ciphertext of the facial features of all employees;

[0011] The homomorphic encryption public key is configured by a trusted institution.

[0012] Optionally, the data processor is also used to store the encrypted text in the corresponding enterprise's local database.

[0013] Optionally, a main chain may be selected from the park's blockchain and all enterprise blockchains to serve as the data synchronization standard, including:

[0014] The hash values ​​of the blockchain in the park and the hash values ​​of all enterprise blockchains are statistically analyzed, and the enterprise blockchain with the most hash values ​​is selected as the main chain.

[0015] Optionally, the encrypted data stored in the main chain can be synchronized to the park's blockchain and all enterprise blockchains other than the main chain, including:

[0016] The threshold private key share is obtained from each enterprise's local database through a smart contract, and all the obtained threshold private key shares are concatenated into a threshold private key; the threshold private key and the threshold public key are a threshold key pair, the threshold public key is used to encrypt each enterprise's blockchain, and the threshold private key share is configured by a trusted institution;

[0017] The main chain is decrypted using a threshold private key via a smart contract, yielding the ciphertext stored in the main chain.

[0018] The encrypted data stored in the main chain is synchronized to the park's blockchain and all enterprise blockchains other than the main chain via smart contracts.

[0019] Optionally, the blockchain module is also used to encrypt each enterprise blockchain using a threshold public key after synchronizing the ciphertext stored in the main chain to the park blockchain and all enterprise blockchains other than the main chain.

[0020] Optionally, the park's facial recognition security system may also include:

[0021] The park identification processor is used to acquire the facial features of people entering the park, encrypt the facial features using a homomorphic encryption public key to obtain the ciphertext of the facial features, and open the park's access control system when the ciphertext of the facial features of the person is present in the ciphertext stored in the park's blockchain.

[0022] Optionally, the park's facial recognition security system may also include:

[0023] The enterprise identification processor is used to acquire the facial features of personnel entering the enterprise corresponding to the enterprise identification processor, and encrypt the facial features of the personnel using a homomorphic encryption public key to obtain the ciphertext of the personnel's facial features. When the ciphertext of the personnel's facial features is found in the ciphertext stored in the enterprise's local database, the enterprise's access control system is opened.

[0024] Optionally, the face recognition model used to obtain facial features in the campus recognition processor is the same as the face recognition model used to obtain facial features in the enterprise recognition processor. The face recognition model is obtained through federated learning. The central server participating in federated learning is located on the campus side, and the multiple clients participating in federated learning are located on the enterprise side of the campus.

[0025] The above-mentioned solution in this application has the following beneficial effects:

[0026] In the embodiments of this application, when an attacker attempts to delete or tamper with the data stored in the park's access control system's blockchain, the smart contract can detect that the hash value of the park's blockchain is inconsistent with the hash value of the enterprise blockchain. At this time, the smart contract will execute a data synchronization process, determine the main chain as the data synchronization standard from the park's blockchain and all enterprise blockchains, and then synchronize the data stored in the main chain to the park's blockchain, thereby achieving automatic data recovery, preventing the access control system from crashing, and ensuring that park employees can pass through normally.

[0027] Furthermore, since both the enterprise blockchain and the park blockchain store encrypted facial features of employees, while the park access system only stores encrypted facial features and not facial information, attackers cannot obtain employees' facial information when the park access system is successfully attacked, thus effectively protecting the personal data security of employees.

[0028] Other beneficial effects of this application will be described in detail in the following detailed description section. Attached Figure Description

[0029] To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0030] Figure 1 A schematic diagram of the structure of a blockchain-based campus facial recognition security system provided in one embodiment of this application;

[0031] Figure 2 This is a flowchart illustrating the process of employees entering the park, as provided in one embodiment of this application. Detailed Implementation

[0032] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0033] It should be understood that, when used in this application specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or a collection thereof.

[0034] It should also be understood that the term “and / or” as used in this application specification and the appended claims means any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0035] As used in this application specification and the appended claims, the term "if" may be interpreted, depending on the context, as "when," "once," "in response to determination," or "in response to detection." Similarly, the phrase "if determined" or "if detected [the described condition or event]" may be interpreted, depending on the context, as meaning "once determined," "in response to determination," "once detected [the described condition or event]," or "in response to detection [the described condition or event]."

[0036] Furthermore, in the description of this application and the appended claims, the terms "first," "second," "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0037] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0038] To address the issues of system paralysis and personal data leakage when a park access system is successfully attacked, this application provides a blockchain-based facial recognition security system for parks. When an attacker attempts to delete or tamper with data stored in the park's blockchain, the smart contract can detect inconsistencies between the hash values ​​of the park's blockchain and those of the enterprise blockchains. In this case, the smart contract executes a data synchronization process, determining a master chain from the park's blockchain and all enterprise blockchains as the data synchronization standard. The data stored in the master chain is then synchronized to the park's blockchain, enabling automatic data recovery, preventing access system paralysis, and ensuring normal access for park employees.

[0039] Furthermore, since both the enterprise blockchain and the park blockchain store encrypted facial features of employees, while the park access system only stores encrypted facial features and not facial information, attackers cannot obtain employees' facial information when the park access system is successfully attacked, thus effectively protecting the personal data security of employees.

[0040] The following describes the blockchain-based campus facial recognition security system provided in this application by way of specific embodiments.

[0041] like Figure 1 As shown, the blockchain-based facial recognition security system for industrial parks provided in this application includes multiple data processors (such as...) that correspond one-to-one with multiple enterprises within the park. Figure 1 The data processors (1, 2 to n, where n is the number of companies in the park) and the blockchain module are included.

[0042] The data processor can be located on the corresponding enterprise side (such as within the enterprise's management server). The data processor is used to obtain encrypted facial features of all employees of the corresponding enterprise and upload the encrypted data to the enterprise's blockchain. For example... Figure 1 As shown, data processor 1 uploads the obtained ciphertext to the corresponding enterprise's blockchain 1, data processor 2 uploads the obtained ciphertext to the corresponding enterprise's blockchain 2, and data processor n uploads the obtained ciphertext to the corresponding enterprise's blockchain n.

[0043] In some embodiments of this application, the data processor is specifically used to: acquire the facial features of all employees of the enterprise corresponding to the data processor, and encrypt each facial feature using a homomorphic encryption public key (configured by a trusted institution) to obtain ciphertext of the facial features of all employees. After obtaining the ciphertext of the facial features of all employees of the corresponding enterprise, the data processor is also used to store the ciphertext in the local database of the corresponding enterprise for facial recognition of personnel entering the enterprise.

[0044] Specifically, the process by which the data processor obtains ciphertext of facial features is as follows: First, facial data of employees within the company is collected. Then, through face detection, alignment, and normalization, a processed image is obtained. Feature extraction is performed on the processed image to obtain a feature vector (which stores key information for face comparison). Subsequently, the feature vector is encrypted using a homomorphic encryption public key to obtain ciphertext.

[0045] The blockchain module is used to synchronize the encrypted data in each enterprise blockchain to other enterprise blockchains through a consensus mechanism (such as the PBFT consensus mechanism), and to synchronize the encrypted data in each enterprise blockchain to the park's blockchain through smart contracts, so that the encrypted data of the facial features of all employees of all enterprises (i.e., all enterprises in the park) are stored in both the enterprise blockchain and the park blockchain.

[0046] In this system, all enterprise blockchains and the park's blockchain are on the same consortium blockchain network. Specifically, a trusted Certificate Authority (CA) can be designated to distribute identities to the park and its member enterprises, set consortium blockchain permissions, and ensure that the entire park and its internal enterprises are on the same consortium blockchain network, generating public-private key pairs and certificates for each participant. Specifically, the trusted authority will generate homomorphic encryption key pairs and threshold key pairs. The homomorphic encryption key pair consists of a homomorphic public key and a homomorphic private key. The trusted authority will distribute the homomorphic public key to each participant in the consortium blockchain network. The park and all enterprises use the same homomorphic public key to encrypt their respective facial features, and the homomorphic public key is stored by the trusted authority. The threshold key pair consists of a threshold private key and a threshold public key. The trusted authority will divide the threshold private key into n portions (n ​​being the number of enterprises in the park) and distribute these n portions to the n enterprises in the consortium blockchain network. The threshold public key is used to encrypt each enterprise's blockchain.

[0047] During the execution of the PBFT consensus mechanism, each enterprise's blockchain takes turns acting as the master node. Using the PBFT consensus mechanism, after five phases—request, pre-prepare, prepare, commit, and reply—the encrypted feature vectors (i.e., ciphertext of facial features) of each enterprise's employees are packaged into their respective enterprise blockchains in batches. It should be noted that when an enterprise experiences employee changes (such as adding new employees), that enterprise's blockchain can initiate a consensus process to synchronize the ciphertext with other blockchains and the park's blockchain through the consensus mechanism, ensuring consistency between the ciphertext stored in all enterprise blockchains (i.e., all enterprises within the park) and the park's blockchain.

[0048] To further protect data and prevent enterprises from storing easily accessible feature vectors of other enterprises' employees, a threshold encryption smart contract is introduced into the PBFT consensus process. A CA (Certificate Authority) organizes and generates threshold private key shares and threshold public keys. Enterprises can store their threshold private key shares in their local databases. When PBFT execution reaches the commit phase, the smart contract obtains the hash value of the current state blockchain and automatically uses the threshold public key to perform threshold encryption on each enterprise's blockchain, replacing the reply phase's function of uploading block information to the chain and releasing all consensus messages from each enterprise during the consensus process. Before the next PBFT consensus pre-prepare phase begins, the smart contract requests each enterprise's threshold private key share to decrypt their local chain, and then executes each phase of PBFT.

[0049] The aforementioned blockchain module is also used to compare the hash value of the park blockchain with the hash value of each enterprise blockchain through a smart contract. If the hash value of the park blockchain is inconsistent with the hash value of any enterprise blockchain, a main chain is determined from the park blockchain and all enterprise blockchains to serve as the data synchronization standard, and the ciphertext stored in the main chain is synchronized to the other chains in the park blockchain and all enterprise blockchains other than the main chain.

[0050] The hash value of the park blockchain is obtained by calculating the ciphertext stored in the park blockchain using a hash algorithm, while the hash value of the enterprise blockchain is obtained by calculating the ciphertext stored in the enterprise blockchain using a hash algorithm. (It should be noted that the hash value of the enterprise blockchain is obtained by calculating the ciphertext stored in the enterprise blockchain after data synchronization is completed during the consensus mechanism execution process, but before the data is encrypted using the threshold public key.) As a preferred example, the smart contract in this application embodiment can be a Heartbeat smart contract. The Heartbeat smart contract can periodically compare the hash values ​​of the current blockchains of all participants in the consortium blockchain. If any data inconsistency occurs, it automatically requests each enterprise to use its threshold private key share to decrypt its local blockchain (i.e., the enterprise blockchain). The smart contract automatically updates the blockchain of the party with inconsistent data to ensure data correctness.

[0051] In some embodiments of this application, the blockchain module specifically determines the main chain to serve as the data synchronization standard from the park blockchain and all enterprise blockchains in the following manner: It statistically analyzes the hash values ​​of the park blockchain and all enterprise blockchains, and selects the enterprise blockchain with the most hash values ​​as the main chain. That is, the chain with the most identical hash values ​​is selected as the main chain, synchronizing inconsistent blockchain data.

[0052] In some embodiments of this application, the blockchain module specifically synchronizes the ciphertext stored in the main chain to the park blockchain and other chains in all enterprise blockchains besides the main chain in the following manner: First, it obtains a threshold private key share from each enterprise's local database through a smart contract, and concatenates all the obtained threshold private key shares into a threshold private key (the threshold private key and the threshold public key form a threshold key pair, the threshold public key is used to encrypt each enterprise blockchain, and the threshold private key share is configured by a trusted institution); then, it decrypts the main chain using the threshold private key through a smart contract to obtain the ciphertext stored in the main chain; finally, it synchronizes the ciphertext stored in the main chain to the park blockchain and other chains in all enterprise blockchains besides the main chain (i.e., chains whose hash values ​​are inconsistent with the main chain and hash values) through a smart contract.

[0053] Understandably, in order to protect data, the blockchain module is also used to encrypt each enterprise blockchain using threshold public keys after synchronizing the ciphertext stored in the main chain to the park blockchain and all enterprise blockchains other than the main chain.

[0054] In some embodiments of this application, the aforementioned campus facial recognition security system further includes: a campus recognition processor and an enterprise recognition processor. The campus recognition processor can be located on the campus side and is used to acquire the facial features of personnel entering the campus, encrypt the facial features using a homomorphic encryption public key to obtain ciphertext of the personnel's facial features, and activate the campus access control system when the ciphertext of the personnel's facial features is present in the ciphertext stored in the campus blockchain. The enterprise recognition processor can be located on the corresponding enterprise side (as a preferred example, one enterprise recognition processor can be set up on each enterprise side), and is used to acquire the facial features of personnel entering the enterprise corresponding to the enterprise recognition processor, encrypt the facial features using a homomorphic encryption public key to obtain ciphertext of the personnel's facial features, and activate the enterprise's access control system when the ciphertext of the personnel's facial features is present in the ciphertext stored in the enterprise's local database.

[0055] The facial recognition model used to acquire facial features in the park recognition processor is the same as the facial recognition model used in the enterprise recognition processor. This facial recognition model is obtained through federated learning. The central server participating in the federated learning is located on the park side, and the multiple clients participating in the federated learning are located on the enterprise sides of the park, corresponding to each enterprise side. That is, the central server sends initial model parameters to each client. Each client uses the data of the corresponding enterprise (i.e., facial feature data, etc.) to train locally and uploads the trained model parameters to the central server in the park. The central server then performs model aggregation to obtain a global model. This process is repeated until the global model converges. The converged global model is used as the aforementioned facial recognition model and distributed to each client so that the park side and each enterprise side can use the same facial recognition model to extract facial features from facial images. It should be noted that the aforementioned facial recognition model can be a commonly used facial recognition model, such as the VGGFace model.

[0056] In another embodiment of this application, the face recognition model described above can also be a pre-trained model, which does not require training using data from park employees.

[0057] In some other embodiments of this application, in order to facilitate employees to quickly enter the company after passing through the park access control, the company's local database can also store the facial images of the company's employees. The company's recognition processor does not need to perform complex encryption calculations on the facial images. It can directly compare the facial images of the employees with the facial images stored in the local database. If a facial image matching the employee's facial image exists in the local database, the company's access control system is activated, which facilitates the employee to quickly enter the company.

[0058] like Figure 2 As shown in the embodiment of this application, the overall process for employees to enter the park using the blockchain-based facial recognition security system is as follows:

[0059] Step 21: Employees upload their personal information to their employer according to the onboarding process;

[0060] Step 22: The employed company performs face detection, alignment, normalization, and feature vector extraction based on the employee's facial information, followed by homomorphic encryption.

[0061] Step 23: The employed enterprise initiates a consensus process to add the encrypted facial features of the employee to the local enterprise blockchain and participates in updating the park blockchain and other enterprise blockchains.

[0062] Step 24: When employees pass through, they first authenticate at the facial recognition terminal in the park. After the employee undergoes facial recognition, the terminal identifies the facial feature vector, encrypts the feature vector using a homomorphic encryption public key, and compares it with the ciphertext in the park's blockchain. If the comparison is successful, the employee is allowed to pass.

[0063] Step 25: When an employee enters the company, they need to be identified at the company's facial recognition terminal. The terminal identifies the facial feature vector, encrypts the feature vector using a homomorphic encryption public key, and compares it with the ciphertext in the company's local database. If the comparison is successful, the employee is allowed to enter.

[0064] In summary, when attackers attempt to delete or tamper with data stored in the park's access control system's blockchain, the smart contract can detect inconsistencies between the hash values ​​of the park's blockchain and the enterprise blockchains. In this case, the smart contract will execute a data synchronization process, determining the main chain as the data synchronization standard from the park's blockchain and all enterprise blockchains, and then synchronizing the data stored in the main chain to the park's blockchain. This achieves automatic data recovery, prevents the access control system from crashing, and ensures that park employees can pass through normally.

[0065] Furthermore, since both the enterprise blockchain and the park blockchain store encrypted facial features of employees, while the park access system only stores encrypted facial features and not facial information, attackers cannot obtain employees' facial information when the park access system is successfully attacked, thus effectively protecting the personal data security of employees.

[0066] The above description is the preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principles described in this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A blockchain-based facial recognition security system for industrial parks, characterized in that, include: Multiple data processors are associated with multiple enterprises within the park. The data processors are used to obtain encrypted facial features of all employees of the enterprise corresponding to the data processor and upload the encrypted data to the enterprise's blockchain. The blockchain module is used to synchronize the encrypted data in each enterprise blockchain to other enterprise blockchains through a consensus mechanism, and to synchronize the encrypted data in each enterprise blockchain to the park blockchain through a smart contract, so that the encrypted data of the facial features of the employees of all enterprises are stored in the enterprise blockchains and the park blockchain; all enterprise blockchains and the park blockchain are in the same consortium blockchain network. The blockchain module is also used to compare the hash value of the park blockchain with the hash value of each enterprise blockchain through a smart contract. If the hash value of the park blockchain is inconsistent with the hash value of any enterprise blockchain, a main chain is determined from the park blockchain and all enterprise blockchains to serve as the data synchronization standard, and the ciphertext stored in the main chain is synchronized to the other chains in the park blockchain and all enterprise blockchains other than the main chain.

2. The campus facial recognition security system according to claim 1, characterized in that, The data processor is specifically used to: acquire the facial features of all employees of the enterprise corresponding to the data processor, and encrypt each facial feature using a homomorphic encryption public key to obtain the ciphertext of the facial features of all employees; The homomorphic encryption public key is configured by a trusted institution.

3. The campus facial recognition security system according to claim 1, characterized in that, The data processor is also used to store the encrypted text in the corresponding enterprise's local database.

4. The campus facial recognition security system according to claim 2, characterized in that, The process of determining the main chain from the park's blockchain and all enterprise blockchains to serve as the data synchronization standard includes: The hash values ​​of the park's blockchain and all enterprise blockchains are statistically analyzed, and the enterprise blockchain with the highest number of hash values ​​is selected as the main chain.

5. The campus facial recognition security system according to claim 4, characterized in that, The step of synchronizing the ciphertext stored in the main chain to the park blockchain and all enterprise blockchains other than the main chain includes: The threshold private key share is obtained from the local database of each of the enterprises through a smart contract, and all the obtained threshold private key shares are concatenated into a threshold private key; the threshold private key and the threshold public key are a threshold key pair, the threshold public key is used to encrypt the blockchain of each of the enterprises, and the threshold private key share is configured by a trusted institution; The main chain is decrypted using the threshold private key via a smart contract to obtain the ciphertext stored in the main chain; The encrypted data stored in the main chain is synchronized to the park blockchain and all enterprise blockchains other than the main chain via smart contracts.

6. The campus facial recognition security system according to claim 5, characterized in that, The blockchain module is further configured to encrypt each of the enterprise blockchains using the threshold public key after synchronizing the ciphertext stored in the main chain to the park blockchain and other chains in all enterprise blockchains besides the main chain.

7. The campus facial recognition security system according to claim 2, characterized in that, The park's facial recognition security system also includes: The park identification processor is used to acquire the facial features of people entering the park, encrypt the facial features of the people using a homomorphic encryption public key to obtain the ciphertext of the facial features of the people, and open the park's access control system when the ciphertext of the facial features of the people is present in the ciphertext stored in the park's blockchain.

8. The campus facial recognition security system according to claim 7, characterized in that, The park's facial recognition security system also includes: The enterprise identification processor is used to acquire the facial features of personnel entering the enterprise corresponding to the enterprise identification processor, encrypt the facial features of the personnel using a homomorphic encryption public key to obtain the ciphertext of the personnel's facial features, and open the enterprise's access control system when the ciphertext of the personnel's facial features is present in the ciphertext stored in the enterprise's local database.

9. The campus facial recognition security system according to claim 8, characterized in that, The face recognition model used to acquire facial features in the park recognition processor is the same as the face recognition model used to acquire facial features in the enterprise recognition processor. The face recognition model is obtained through federated learning. The central server participating in the federated learning is located on the park side, and the multiple clients participating in the federated learning are located on the enterprise side of the park.

Images