Information collection and management system based on face recognition technology

Abstract

The present invention discloses an information collection and management system based on face recognition technology, and relates to the field of face recognition technology. The system comprises the following components: a face recognition module, which is responsible for acquiring face images through a camera or a picture acquisition device, and performing face detection and recognition using a deep learning algorithm. During the recognition process, the system uses differential privacy technology to perform perturbation processing on the original face images to protect user privacy. The original face images are perturbed by the differential privacy technology in the face recognition module, making it extremely difficult to extract personal privacy information from these data. Secondly, in terms of data security, the present invention uses an efficient encryption algorithm in a data encryption module to encrypt the collected face images and recognition results, thereby ensuring the confidentiality and integrity of the data during transmission and storage, and effectively preventing the security risks of data leakage and tampering.

Description

Technical Field

[0001] The present invention relates to the field of face recognition technology, in particular to an information collection and management system based on face recognition technology. Background Art

[0002] With the rapid development of information technology, face recognition technology has penetrated into all aspects of our daily lives and has become a key link in modern social management and security. However, the widespread application of this technology is also accompanied by a series of profound privacy and security issues. As the core part of personal biometric information, facial images are unique and non-replicable, making them a high-risk point for privacy leakage. Once these image data are illegally obtained, abused or leaked, personal privacy will be seriously violated, which may lead to a series of adverse consequences such as identity impersonation and fraud.

[0003] Therefore, how to protect user privacy and improve system security while ensuring the efficient application of face recognition technology is a problem that needs to be solved urgently. At present, some technical means have been used to solve these problems, such as differential privacy technology and encryption technology. However, these technical means still have some challenges and limitations in practical applications. How to maintain recognition accuracy while protecting privacy, how to ensure the security and availability of encrypted data and other potential threats constitute the core technical solution of the present invention, which provides strong support for the privacy protection and security enhancement of face recognition technology.

[0004] However, the information collection and management systems based on face recognition technology on the market may face security risks of data leakage, tampering and illegal access when processing face image data, and cannot achieve user privacy protection and system security. Therefore, the development of an information collection and management system based on face recognition technology that can comprehensively realize the above-mentioned characteristics provides strong support for the privacy protection and security enhancement of face recognition technology. Summary of the invention

[0005] The purpose of this invention is to make up for the shortcomings of the existing technology and provide an information collection and management system based on face recognition technology. It can ensure the security of data during transmission and storage by adopting advanced encryption technology, differential privacy technology and security audit mechanism, and timely discover and deal with potential security risks, thereby providing strong support for the privacy protection and security enhancement of face recognition technology.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: an information collection and management system based on face recognition technology, the system includes the following components:

[0007] Face recognition module: This module is responsible for acquiring face images through a camera or image acquisition device, and using a deep learning algorithm to detect and recognize faces. During the recognition process, the system uses differential privacy technology to perturb the original face image to protect user privacy.

[0008] Data encryption module: This module is responsible for encrypting the collected face images and recognition results. The system uses an efficient encryption algorithm to ensure the security of data during transmission and storage;

[0009] Database management module: This module is responsible for storing and managing encrypted face images and recognition results. The database adopts an access control mechanism, and only authorized users can access relevant data;

[0010] Security audit module: This module is responsible for real-time monitoring and auditing of the system's security. The system records user operation logs and detects abnormal behavior to promptly identify and handle potential security risks.

[0011] Furthermore, the specific steps of the face recognition module are:

[0012] (1) Image acquisition and preprocessing: The user's face image is obtained through a camera or image acquisition device. In order to improve the accuracy of subsequent face detection and recognition, the system will preprocess the acquired image, including adjusting the image size, grayscale, noise reduction, and contrast enhancement operations to eliminate background noise and uneven lighting interference factors;

[0013] (2) Face detection: The preprocessed image will enter the face detection stage. This detection uses the convolutional neural network (CNN) in the deep learning algorithm to build an efficient face detector that can accurately and quickly locate the face area in a complex background. The detector will output the location information of the face, including coordinates and size, to provide a basis for subsequent feature extraction and recognition.

[0014] (3) Feature extraction and recognition: After detecting a face, the system extracts features from the face area. These features can be texture, shape, and depth information, which can characterize the uniqueness of the face. After feature extraction is completed, the system compares the extracted features with the features stored in the database to determine the identity of the face.

[0015] (4) Differential privacy protection: In order to protect user privacy, this module introduces differential privacy technology after feature extraction. By adding random noise or other forms of perturbations to the original feature vector, the change of a single user's data has little effect on the overall recognition result. The implementation of differential privacy needs to consider the privacy budget and noise addition strategy to strike a balance between protecting privacy and maintaining recognition performance.

[0016] Furthermore, the data encryption module comprises the following specific steps:

[0017] (1) Data preparation: The data encryption module receives data from the face recognition module. The data mainly includes two parts: face image and recognition result. The face image is the original image data, which contains the user's personal features. The recognition result is the user's identity information recognized by the system based on the face image.

[0018] (2) Key generation: Key generation is a critical step in the encryption process. According to the system's security policy and the requirements of the asymmetric encryption algorithm, the key generation algorithm will generate one or more keys. These keys are used in subsequent encryption operations to ensure the confidentiality of the data.

[0019] (3) Data encryption storage: Use the selected encryption algorithm and generated key to encrypt the face image and recognition results, and store the encrypted data in the database management module to ensure that only authorized users can access it.

[0020] Furthermore, the database management module specifically comprises the following steps:

[0021] (1) Data reception and storage: First, the encrypted face image and recognition result data are received from the data encryption module, and the received encrypted data is stored in the database using partition storage and distributed storage technology;

[0022] (2) Access control: The database management module implements a strict access control mechanism, with built-in machine learning-based abnormal loss function detection, user authentication, permission verification, and access log recording steps when access is required, ensuring that only authorized users can access the data stored in the database;

[0023] (3) Data backup and recovery: In order to prevent data loss or damage, the database management module will regularly back up the stored data. When data loss or damage occurs, the backup data can be used for recovery to ensure the normal operation of the system and the security of the data;

[0024] (4) Data retrieval and response: When an authorized user initiates a data retrieval request, the database management module will retrieve the corresponding encrypted data from the database based on the user's permissions and request content, and return it to the user. At the same time, the module will record the user's retrieval behavior for subsequent security audits and analysis.

[0025] Furthermore, the security audit module specifically includes the following steps:

[0026] (1) Log collection and analysis: The security audit module first collects log information from each module in the system, including user operation logs, system event logs, and exception logs, parses and analyzes the collected logs, and extracts key information, such as user behavior patterns, access time, and access frequency;

[0027] (2) Risk detection: Based on the results of log analysis, use the preset security rules and mean and standard deviation algorithms to detect whether there are abnormal behaviors or potential security risks in the system;

[0028] (3) Alert and response: If abnormal behavior or security risks are detected, the security audit module will trigger the alert mechanism, notify the system administrator or relevant security personnel, and take appropriate response measures, such as blocking access and recording evidence;

[0029] (4) Report generation: Generate security audit reports regularly based on demand, summarize and evaluate the security of the system, and provide decision support for security management.

[0030] Furthermore, the face recognition module can effectively extract image features through the convolutional neural network CNN in the deep learning algorithm, thereby realizing efficient and accurate face recognition. The mathematical formula of the convolutional layer is: S(i,j)=(I*K)(i,j)=∑ m ∑ n I(i+m,j+n)K(m,n), I is the input image, K is the convolution kernel, S is the result after convolution, the convolution layer is the core component of CNN, it performs convolution operation on the input face image through multiple convolution kernels to extract local features in the image, these convolution kernels can be regarded as filters, they can automatically learn and recognize the basic information of edges and textures in the image, with the increase of the number of network layers, the convolution layer can extract more advanced and abstract features, such as the contour, eyes, and nose of the face, the application of CNN algorithm enables the system to automatically learn and extract key facial features from the original image without complex manual feature extraction, which not only improves the accuracy and efficiency of face recognition, but also enables the system to adapt to different scenes, different lighting conditions, and different complex situations of face postures.

[0031] Furthermore, the face recognition module performs differential privacy protection on the face features extracted from the face image through differential privacy processing. Suppose the original feature vector is x, the privacy budget is, and the noise addition strategy is to add random noise n that follows a certain distribution to each element of the feature vector. Its mathematical expression is: x ′=x+n, the differential privacy mechanism will add noise to these extracted features to achieve the purpose of protecting privacy. The added noise obeys a specific distribution, and its intensity is related to the privacy budget. The smaller the value, the more noise is added, and the higher the degree of privacy protection, but the corresponding data utility will also be reduced. Therefore, through differential privacy processing, the face recognition system can protect personal privacy while still maintaining effective recognition of the face. This not only meets people's needs for privacy protection, but also gives play to the advantages of face recognition technology in identity authentication and security monitoring.

[0032] Furthermore, the data encryption module uses an asymmetric encryption algorithm (RSA), also known as a public key encryption algorithm, which usually uses a pair of public and private keys. The public key is used to encrypt data, while the private key is used to decrypt data. The main feature of this encryption method is that the public key can be shared publicly, while the private key must be kept confidential. Its asymmetric encryption algorithm first generates a key by selecting two large prime numbers p and q, calculating ln=p×q, selecting an integer e that is relatively prime to φ(n), where φ(n)=(p-1)×(q-1), which is the Euler function value of n, calculating d, so that d×emodφ(n)=1, the public key is (n,e), and the private key is (n,d), and then encrypting the key to convert the plaintext message m into an integer less than n. The encryption process is: c=m Λ e mod n, for ciphertext c, the decryption process is: m = c Λ d mod n, m is the decrypted plaintext. Since only users with private keys can decrypt data, public keys can be securely stored and transmitted. Therefore, even if the data is intercepted during transmission, attackers cannot easily obtain plaintext information, which ensures the confidentiality of face images and recognition results.

[0033] Furthermore, the database management module uses the loss function to detect the mathematical formula: y i is the true value, is the predicted value, and n is the number of samples. In the model training process of machine learning and deep learning, it is used to measure the gap between the model predicted value and the true value, and optimize the model parameters accordingly. When training a machine learning model, it is necessary to obtain training data from the database and use the loss function to guide the model parameter update. As the model training progresses, the loss function will drive the optimization of the model parameters, thereby indirectly affecting the way the training data stored in the database is used.

[0034] Furthermore, the security audit module calculates the mean and standard deviation of data of different time periods, different users or different system components through the mean and standard deviation algorithm, so as to understand the overall distribution and fluctuation of the data and discover abnormal patterns or trends. The mean provides the center point of the data, and its mathematical expression is: The standard deviation describes the degree of dispersion of the data, and its expression is: If the value of a data point deviates significantly from the mean of its data set, it may be considered an outlier. These outliers may represent potential security issues, such as abnormal login attempts, abnormal traffic, or abnormal behavior patterns. By comparing the means and standard deviations of different data sets, different users, and different time periods, the risk level in different scenarios can be assessed. If security risks or loopholes are found, the system will issue an alarm in a timely manner and notify the administrator or relevant personnel to take appropriate measures, such as strengthening system security and restricting user access.

[0035] Compared with the existing technology, the information collection and management system based on face recognition technology has the following beneficial effects:

[0036] First, the present invention uses the differential privacy technology in the face recognition module to perturb the original face image, making any attempt to extract personal privacy information from these data extremely difficult. Secondly, in terms of data security, the present invention uses an efficient encryption algorithm in the data encryption module to encrypt the collected face images and recognition results, ensuring the confidentiality and integrity of the data during transmission and storage, and effectively preventing security risks such as data leakage and tampering.

[0037] 2. In terms of system security, the present invention realizes comprehensive monitoring and management of system security through database management and security audit modules. This all-round security monitoring mechanism greatly enhances the stability and security of the system and provides users with more reliable and secure services.

[0038] Other advantages, objectives and features of the present invention will be set forth in part in the following description and, in part, will be apparent to those skilled in the art based on an examination of the following or may be taught from the practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0040] Figure 1 This is the process operation diagram of the information collection and management system based on face recognition technology.

[0041] Figure 2 This is a flow chart of the information collection and management system based on face recognition technology. DETAILED DESCRIPTION

[0042] The technical solutions in the embodiments of the present invention are described clearly and completely below. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. 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.

[0043] Example 1

[0044] The information collection and management system based on face recognition technology includes the following components:

[0045] Face recognition module: This module is responsible for acquiring face images through a camera or image acquisition device, and using a deep learning algorithm to detect and recognize faces. During the recognition process, the system uses differential privacy technology to perturb the original face image to protect user privacy.

[0046] Data encryption module: This module is responsible for encrypting the collected face images and recognition results. The system uses an efficient encryption algorithm to ensure the security of data during transmission and storage;

[0047] Database management module: This module is responsible for storing and managing encrypted face images and recognition results. The database adopts an access control mechanism, and only authorized users can access relevant data;

[0048] Security audit module: This module is responsible for real-time monitoring and auditing of the system's security. The system records user operation logs and detects abnormal behavior to promptly identify and handle potential security risks.

[0049] The data encryption module uses an asymmetric encryption algorithm (RSA), also known as a public key encryption algorithm. It usually uses a pair of public and private keys. The public key is used to encrypt data, while the private key is used to decrypt data. The main feature of this encryption method is that the public key can be shared publicly, while the private key must be kept confidential. Its asymmetric encryption algorithm first generates a key by selecting two large prime numbers p and q, calculating ln=p×q, and selecting an integer e that is relatively prime to φ(n), where φ(n)=(p-1)×(q-1), which is the Euler function value of n. Calculate d so that d×emodφ(n)=1, the public key is (n,e), and the private key is (n,d). Then encrypt the key to convert the plaintext message m into an integer less than n. The encryption process is: c=m Λ e mod n, for ciphertext c, the decryption process is: m = c Λd mod n, m is the decrypted plaintext. Since only users with private keys can decrypt data, public keys can be securely stored and transmitted. Therefore, even if the data is intercepted during transmission, attackers cannot easily obtain plaintext information, which ensures the confidentiality of face images and recognition results.

[0050] The collected facial images and recognition results are encrypted through the encryption algorithm adopted by the data encryption module. During the encryption process, the module will encrypt the original data to generate encrypted data. Even if the encrypted data is illegally obtained during transmission and storage, the attacker cannot easily decrypt and obtain the sensitive information therein, ensuring the security of data collection, transmission and storage, and preventing the data from being illegally obtained or tampered with.

[0051] Example 2

[0052] The system consists of the following components:

[0053] Face recognition module: This module is responsible for acquiring face images through a camera or image acquisition device, and using a deep learning algorithm to detect and recognize faces. During the recognition process, the system uses differential privacy technology to perturb the original face image to protect user privacy.

[0054] Data encryption module: This module is responsible for encrypting the collected face images and recognition results. The system uses an efficient encryption algorithm to ensure the security of data during transmission and storage;

[0055] Database management module: This module is responsible for storing and managing encrypted face images and recognition results. The database adopts an access control mechanism, and only authorized users can access relevant data;

[0056] Security audit module: This module is responsible for real-time monitoring and auditing of the system's security. The system records user operation logs and detects abnormal behavior to promptly identify and handle potential security risks.

[0057] The security audit module calculates the mean and standard deviation of data from different time periods, different users or different system components through the mean and standard deviation algorithm, which can understand the overall distribution and fluctuation of the data and thus discover abnormal patterns or trends. The mean provides the center point of the data, and its mathematical expression is: The standard deviation describes the degree of dispersion of the data, and its expression is: If the value of a data point deviates significantly from the mean of its data set, it may be considered an outlier. These outliers may represent potential security issues, such as abnormal login attempts, abnormal traffic, or abnormal behavior patterns. By comparing the means and standard deviations of different data sets, different users, and different time periods, the risk level in different scenarios can be assessed. If security risks or loopholes are found, the system will issue an alarm in a timely manner and notify the administrator or relevant personnel to take appropriate measures, such as strengthening system security and restricting user access.

[0058] The security audit module uses the mean and standard deviation algorithms to conduct comprehensive security checks on the system. These tests include vulnerability scanning, malicious code detection, and intrusion detection, which aim to discover security risks and vulnerabilities in the system and take corresponding measures in a timely manner. The security of the system is monitored and audited in real time, potential security risks are discovered and handled in a timely manner, and the protection level of user privacy and system security is effectively improved.

[0059] Example 3

[0060] The steps of the system based on Example 1 and Example 2 are as follows:

[0061] (1) Facial image acquisition and recognition

[0062] Users upload facial images through cameras or image acquisition devices;

[0063] The face recognition module uses deep learning algorithms to detect and recognize faces in uploaded images;

[0064] During the recognition process, the system uses differential privacy technology to perturb the original facial image to ensure user privacy.

[0065] (2) Data encryption

[0066] The data encryption module uses the efficient asymmetric encryption algorithm RSA to encrypt the collected face images and recognition results;

[0067] During the encryption process, the module converts the original data into encrypted data to ensure data security during transmission and storage;

[0068] Even if the encrypted data in the encryption algorithm is illegally obtained, the attacker cannot easily decrypt and obtain the sensitive information contained therein.

[0069] (3) Data storage and management

[0070] The encrypted face image and recognition results are stored in the database management module;

[0071] If a user needs to access encrypted data in the database, the system will verify and authorize based on the preset access control mechanism;

[0072] Only authenticated and authorized users can access and operate relevant data.

[0073] (4) Real-time monitoring

[0074] The security audit module monitors the security of the system in real time and records user operation logs;

[0075] By detecting abnormal behaviors, potential security risks can be discovered and handled in a timely manner;

[0076] The security audit module uses mean and standard deviation algorithms to analyze data from different time periods, different users, or different system components to find abnormal patterns or trends;

[0077] If a data point deviates significantly from the mean, it may be considered an outlier and require further investigation and treatment.

[0078] (5) Security audit and risk management

[0079] The security audit module uses mean and standard deviation algorithms to conduct comprehensive security testing of the system, including vulnerability scanning, malicious code detection, and intrusion detection;

[0080] If security risks or loopholes are found, the system will promptly issue an alarm and notify the administrator or relevant personnel;

[0081] Administrators or relevant personnel take corresponding measures based on the alarm information, such as strengthening system security and restricting user access.

[0082] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

Claims

1. The information collection and management system based on face recognition technology is characterized by: The system consists of the following components: Face recognition module: This module is responsible for acquiring face images through a camera or image acquisition device, and using a deep learning algorithm to detect and recognize faces. During the recognition process, the system uses differential privacy technology to perturb the original face image to protect user privacy. Data encryption module: This module is responsible for encrypting the collected facial images and recognition results. The system uses an efficient encryption algorithm to ensure the security of data during transmission and storage; Database management module: This module is responsible for storing and managing encrypted face images and recognition results. The database adopts an access control mechanism, and only authorized users can access relevant data; Security audit module: This module is responsible for real-time monitoring and auditing of the system's security. The system records user operation logs and detects abnormal behavior to promptly identify and handle potential security risks.

2. The information collection and management system based on face recognition technology according to claim 1 is characterized in that: The specific steps of the face recognition module are: (1) Image acquisition and preprocessing: The user's face image is obtained through a camera or image acquisition device. In order to improve the accuracy of subsequent face detection and recognition, the system will preprocess the acquired image, including adjusting the image size, grayscale, noise reduction, and contrast enhancement operations to eliminate background noise and uneven lighting interference factors; (2) Face detection: The preprocessed image will enter the face detection stage. This detection uses the convolutional neural network (CNN) in the deep learning algorithm to build an efficient face detector that can accurately and quickly locate the face area in a complex background. The detector will output the location information of the face, including coordinates and size, to provide a basis for subsequent feature extraction and recognition. (3) Feature extraction and recognition: After detecting a face, the system extracts features from the face area. These features can be texture, shape, and depth information, which can characterize the uniqueness of the face. After feature extraction is completed, the system compares the extracted features with the features stored in the database to determine the identity of the face. (4) Differential privacy protection: In order to protect user privacy, this module introduces differential privacy technology after feature extraction. By adding random noise or other forms of perturbations to the original feature vector, the change of a single user's data has little effect on the overall recognition result. The implementation of differential privacy needs to consider the privacy budget and noise addition strategy to strike a balance between protecting privacy and maintaining recognition performance.

3. The information collection and management system based on face recognition technology according to claim 1 is characterized in that: The data encryption module has the following specific steps: (1) Data preparation: The data encryption module receives data from the face recognition module. The data mainly includes two parts: face image and recognition result. The face image is the original image data, which contains the user's personal features. The recognition result is the user's identity information recognized by the system based on the face image. (2) Key generation: Key generation is a critical step in the encryption process. According to the system's security policy and the requirements of the asymmetric encryption algorithm, the key generation algorithm will generate one or more keys. These keys are used in subsequent encryption operations to ensure the confidentiality of the data. (3) Data encryption storage: Use the selected encryption algorithm and generated key to encrypt the face image and recognition results, and store the encrypted data in the database management module to ensure that only authorized users can access it.

4. The information collection and management system based on face recognition technology according to claim 1 is characterized in that: The database management module has the following specific steps: (1) Data reception and storage: First, the encrypted face image and recognition result data are received from the data encryption module, and the received encrypted data is stored in the database using partition storage and distributed storage technology; (2) Access control: The database management module implements a strict access control mechanism, with built-in machine learning-based abnormal loss function detection, user authentication, permission verification, and access log recording steps when access is required, ensuring that only authorized users can access the data stored in the database; (3) Data backup and recovery: In order to prevent data loss or damage, the database management module will regularly back up the stored data. When data loss or damage occurs, the backup data can be used for recovery to ensure the normal operation of the system and the security of the data; (4) Data retrieval and response: When an authorized user initiates a data retrieval request, the database management module will retrieve the corresponding encrypted data from the database based on the user's permissions and request content, and return it to the user. At the same time, the module will record the user's retrieval behavior for subsequent security audits and analysis.

5. The information collection and management system based on face recognition technology according to claim 1 is characterized in that: The specific steps of the security audit module are as follows: (1) Log collection and analysis: The security audit module first collects log information from each module in the system, including user operation logs, system event logs, and exception logs, parses and analyzes the collected logs, and extracts key information, such as user behavior patterns, access time, and access frequency; (2) Risk detection: Based on the results of log analysis, use the preset security rules and mean and standard deviation algorithms to detect whether there are abnormal behaviors or potential security risks in the system; (3) Alert and response: If abnormal behavior or security risks are detected, the security audit module will trigger the alert mechanism, notify the system administrator or relevant security personnel, and take appropriate response measures, such as blocking access and recording evidence; (4) Report generation: Generate security audit reports regularly based on demand, summarize and evaluate the security of the system, and provide decision support for security management.

6. The information collection and management system based on face recognition technology according to claim 2 is characterized in that ,The face recognition module can effectively extract image features through the convolutional neural network CNN in the deep learning algorithm, thereby realizing efficient and accurate face recognition. The mathematical formula of its convolutional layer is: S(i,j)=(I*K)(i,j)=∑ m ∑ n I(i+m,j+n)K(m,n), I is the input image, K is the convolution kernel, S is the result after convolution, the convolution layer is the core component of CNN, it performs convolution operation on the input face image through multiple convolution kernels to extract local features in the image, these convolution kernels can be regarded as filters, they can automatically learn and recognize the basic information of edges and textures in the image, with the increase of the number of network layers, the convolution layer can extract more advanced and abstract features, such as the contour, eyes, and nose of the face, the application of CNN algorithm enables the system to automatically learn and extract key facial features from the original image without complex manual feature extraction, which not only improves the accuracy and efficiency of face recognition, but also enables the system to adapt to different scenes, different lighting conditions, and different complex situations of face postures.

7. The information collection and management system based on face recognition technology according to claim 2 is characterized in that The face recognition module performs differential privacy protection on the face features extracted from the face image through differential privacy processing. Let the original feature vector be x, the privacy budget be, and the noise adding strategy be to add random noise n that obeys a certain distribution to each element of the feature vector. Its mathematical expression is: ′ =x+n, the differential privacy mechanism will add noise to these extracted features to achieve the purpose of protecting privacy. The added noise obeys a specific distribution, and its intensity is related to the privacy budget. The smaller the value, the more noise is added, and the higher the degree of privacy protection, but the corresponding data utility will also be reduced. Therefore, through differential privacy processing, the face recognition system can protect personal privacy while still maintaining effective recognition of the face. This not only meets people's needs for privacy protection, but also gives play to the advantages of face recognition technology in identity authentication and security monitoring.

8. The information collection and management system based on face recognition technology according to claim 3 is characterized in that The data encryption module uses an asymmetric encryption algorithm (RSA), also known as a public key encryption algorithm. It usually uses a pair of public and private keys. The public key is used to encrypt data, while the private key is used to decrypt data. The main feature of this encryption method is that the public key can be shared publicly, while the private key must be kept confidential. The asymmetric encryption algorithm first generates a key by selecting two large prime numbers p and q, calculating ln=p×q, and selecting an integer e that is relatively prime to φ(n), where φ(n)=(p-1)×(q-1), which is the Euler function value of n. Calculate d so that d×emodφ(n)=1, the public key is (n,e), and the private key is (n,d). Then encrypt the key to convert the plaintext message m into an integer less than n. The encryption process is: c=m Λ e mod n, for ciphertext c, the decryption process is: m = c Λ d mod n, m is the decrypted plaintext. Since only users with private keys can decrypt data, public keys can be securely stored and transmitted. Therefore, even if the data is intercepted during transmission, attackers cannot easily obtain plaintext information, which ensures the confidentiality of face images and recognition results.

9. The information collection and management system based on face recognition technology according to claim 4 is characterized in that ,The database management module uses the loss function to detect its mathematical formula: y i is the true value, is the predicted value, and n is the number of samples. In the model training process of machine learning and deep learning, it is used to measure the gap between the model predicted value and the true value, and optimize the model parameters accordingly. When training a machine learning model, it is necessary to obtain training data from the database and use the loss function to guide the model parameter update. As the model training progresses, the loss function will drive the optimization of the model parameters, thereby indirectly affecting the way the training data stored in the database is used.

10. The information collection and management system based on face recognition technology according to claim 5, characterized in that The security audit module calculates the mean and standard deviation of data in different time periods, different users or different system components through the mean and standard deviation algorithm, which can understand the overall distribution and fluctuation of the data, so as to find abnormal patterns or trends. The mean provides the center point of the data, and its mathematical expression is: The standard deviation describes the degree of dispersion of the data, and its expression is: If the value of a data point deviates significantly from the mean of its data set, it may be considered an outlier. These outliers may represent potential security issues, such as abnormal login attempts, abnormal traffic, or abnormal behavior patterns. By comparing the means and standard deviations of different data sets, different users, and different time periods, the risk level in different scenarios can be assessed. If security risks or loopholes are found, the system will issue an alarm in a timely manner and notify the administrator or relevant personnel to take appropriate measures, such as strengthening system security and restricting user access.

Images