An intelligent identification analysis and safety protection system for fire identification examinations

Abstract

The present application belongs to the field of data processing, and in particular relates to an intelligent identification analysis and security protection system for fire identification examinations; wherein, the access subject collects real-time monitoring image data of the real-time monitoring object; the real-time monitoring image is input into the identity encryption model to generate a real-time login request adapted to the encryption channel; the real-time login request is sent through the encryption channel; the application trust center uses the identity decryption model to perform intelligent analysis on the received real-time login request to obtain real-time identity information; the real-time identity information is subjected to monitoring behavior analysis to obtain the examination behavior; the real-time identity information is subjected to security risk assessment using a security protection model to obtain a security protection assessment associated with the real-time monitoring object; the real-time monitoring object is marked with the corresponding examination behavior and security protection assessment. The system performs identity identification, behavior analysis and security protection assessment on the fire identification examination process, thereby improving the efficiency of examination supervision and the safety of the examination room.

Description

Technical Field

[0001] The present application belongs to the field of data processing, and in particular relates to an intelligent identity recognition analysis and safety protection system for fire identification examinations. Background Art

[0002] The Fire Protection Certification Examination (FCE) is an assessment of fire safety knowledge, skills, and emergency response capabilities for firefighters. It typically consists of both theoretical knowledge and practical skills assessments, covering knowledge and skills in fire prevention and control, emergency evacuation, and firefighting. Passing the FCE examination directly impacts a firefighter's competence and responsibility in fire safety work, and therefore holds significant practical significance.

[0003] During the exam, the possibility of various anomalies and security issues must be fully considered, as they could seriously threaten the fairness and security of the exam. For example, cheating may occur through plagiarism or the use of communication devices, which directly affects the fairness of the exam and harms the rights of legitimate candidates. For example, monitoring equipment or systems may not function properly due to technical failures, making it impossible to monitor the exam process in real time, which makes it more difficult to detect anomalies during the exam.

[0004] Therefore, in order to ensure the smooth progress and security of the examination process, it is urgent to establish a real-time monitoring system to solve at least one of the above technical problems. Summary of the Invention

[0005] This application provides an intelligent identity recognition, analysis, and security protection system for fire protection examinations. It is used to perform identity recognition, behavior analysis, and security protection assessments on real-time monitored subjects during fire protection examinations, helping to improve the security, fairness, and accuracy of the examination process, as well as the efficiency of examination invigilation and the safety of examination rooms. The system can monitor abnormal situations during the examination in real time and can promptly identify and take necessary measures to respond, ensuring that the fairness and security of the examination are not threatened.

[0006] In a first aspect, the present application provides a system for intelligent identification analysis and security protection of fire protection examination identities, the system comprising an access subject and an application trust center, an encrypted channel being provided between the access subject and the application trust center; wherein,

[0007] An access subject is configured to collect real-time monitoring image data of a real-time monitoring object; input the real-time monitoring image data into an identity encryption model to generate a real-time login request adapted for an encrypted channel; wherein the real-time login request includes at least: user identity information, authentication credentials, and behavioral image information that have been subjected to multiple encryption processes; and transmit the real-time login request to an application trust center via an encrypted channel;

[0008] An application trust center is used to use an identity decryption model to perform intelligent analysis on received real-time login requests to obtain real-time identity information of the real-time monitored object; the real-time identity information includes at least: user identity information, authentication credentials, and behavioral image information obtained through multiple decryption processes; monitoring behavior analysis is performed on the real-time identity information to obtain the examination behavior of the real-time monitored object; a security protection model is used to perform a security risk assessment on the real-time identity information to obtain a security protection assessment associated with the real-time monitored object; corresponding examination behaviors and security protection assessments are marked for the real-time monitored object to complete real-time monitoring of the real-time monitored object.

[0009] In a second aspect, an embodiment of the present application provides a method for intelligent identification, analysis, and security protection of fire identification examination identities. The method is applied to an access subject in a fire identification examination intelligent identification, analysis, and security protection system. The system also includes an application trust center, and an encrypted channel is provided between the access subject and the application trust center. The method includes:

[0010] Collect real-time monitoring image data of real-time monitoring objects;

[0011] Inputting the real-time monitoring image into the identity encryption model to generate a real-time login request adapted to the encryption channel; wherein the real-time login request includes at least: user identity information, authentication credentials, and behavioral image information that have been subjected to multiple encryption processes;

[0012] The real-time login request is sent to the application trust center through an encrypted channel, so that the application trust center can perform monitoring behavior analysis on the real-time identity information to obtain the examination behavior of the real-time monitoring object, and use a security protection model to perform a security risk assessment on the real-time identity information to obtain a security protection assessment associated with the real-time monitoring object, thereby marking the corresponding examination behavior and security protection assessment for the real-time monitoring object to complete real-time monitoring of the real-time monitoring object.

[0013] In a third aspect, an embodiment of the present application provides a method for intelligent identification, analysis, and security protection of fire identification examination identities. The method is applied to an application trust center in a fire identification examination intelligent identification, analysis, and security protection system. The system also includes an access subject, and an encrypted channel is set between the access subject and the application trust center. The method includes:

[0014] Receiving a real-time login request from a real-time monitored object through an encrypted channel; the real-time login request includes at least: user identity information, authentication credentials, and behavioral image information that have been subjected to multiple encryption processes; the real-time login request generates request data adapted to the encrypted channel through an identity encryption model;

[0015] An identity decryption model is used to intelligently analyze the received real-time login request to obtain the real-time identity information of the real-time monitored object; the real-time identity information includes at least: user identity information, authentication credentials, and behavioral image information obtained through multiple decryption processes;

[0016] Performing monitoring behavior analysis on the real-time identity information to obtain the examination behavior of the real-time monitored subject;

[0017] Using a security protection model, a security risk assessment is performed on the real-time identity information to obtain a security protection assessment associated with the real-time monitoring object;

[0018] The corresponding examination behaviors and security protection assessments are marked for the real-time monitoring objects to complete the real-time monitoring of the real-time monitoring objects.

[0019] In a fourth aspect, an embodiment of the present application provides an electronic device, comprising:

[0020] at least one processor, memory, and input-output unit;

[0021] The memory is used to store computer programs, and the processor is used to call the computer programs stored in the memory to execute the first aspect of the intelligent identity recognition analysis and safety protection system for fire identification examinations.

[0022] In a fifth aspect, a computer-readable storage medium is provided, which includes instructions. When the instructions are run on a computer, the computer executes the first aspect of the intelligent identity recognition analysis and safety protection system for fire identification examinations.

[0023] The technical solution provided by the embodiment of the present application provides an intelligent identification analysis and security protection system for fire protection examinations. The system includes an access subject and an application trust center, and an encrypted channel is set between the access subject and the application trust center. The access subject is used to collect real-time monitoring image data of the real-time monitoring object; input the real-time monitoring image into the identity encryption model to generate a real-time login request adapted to the encryption channel. The real-time login request includes at least: user identity information, authentication credentials, and behavioral image information that have been subjected to multiple encryption processes; and the real-time login request is sent to the application trust center through an encrypted channel. Thus, by collecting and analyzing the real-time monitoring image data of the monitoring object in real time, and applying the identity encryption model and decryption model to process the identity information, it is possible to realize intelligent identification and verification of the identity of the real-time monitoring object, and ensure the accuracy and real-time nature of the identity of the monitoring object. The application trust center is configured to intelligently analyze received real-time login requests using an identity decryption model to obtain real-time identity information of the monitored subject. The real-time identity information includes at least user identity information, authentication credentials, and behavioral image information obtained through multiple decryption processes. Monitoring behavior analysis is performed on the real-time identity information to obtain the exam behavior of the monitored subject. The application trust center then intelligently analyzes the real-time login requests using the identity decryption model to obtain the identity and exam behavior information of the monitored subject. This analysis can help identify abnormal behavior or cheating, ensuring the fairness and accuracy of the exam. A security risk assessment is then performed on the real-time identity information using a security protection model to obtain a security protection assessment associated with the monitored subject. By applying the security protection model to the security risk assessment of the real-time identity information, potential security threats and risks can be promptly identified, helping to provide effective security protection measures for the monitored subject and ensuring the security of the exam process. The corresponding exam behavior and security protection assessment are then labeled for the monitored subject to enable real-time monitoring of the monitored subject. This enables timely detection and recording of important events, providing a basis for subsequent processing and decision-making.

[0024] This application's technical solution performs identity recognition, behavioral analysis, and safety assessment on real-time monitored subjects during fire safety certification exams, helping to improve the security, fairness, and accuracy of the exam process, as well as the efficiency of exam proctoring and the safety of the exam room. The system can monitor abnormalities during the exam in real time and promptly identify and take necessary measures to address them, ensuring that the fairness and security of the exam are not threatened. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0026] Figure 1 This is a structural diagram of an intelligent identification analysis and safety protection system for fire protection examinations according to an embodiment of the present application;

[0027] Figure 2 This is a schematic diagram of the login interface of the fire protection appraisal examination in an embodiment of the present application;

[0028] Figure 3 This is a schematic diagram of the test answering interface of the fire protection appraisal test in an embodiment of the present application;

[0029] Figure 4 This is a flow chart of a method for intelligent identification analysis and safety protection of fire protection examination identity in an embodiment of the present application;

[0030] Figure 5 This is a flow chart of another method for intelligent identification analysis and safety protection of fire protection examinations according to an embodiment of the present application;

[0031] Figure 6 It is a structural diagram of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0032] To make the purpose, technical solutions, and advantages of the embodiments of this application more clear, the technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the drawings in the embodiments of this application. Obviously, the described embodiments are part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0033] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which this application pertains. The terms used herein in the specification of this application are for the purpose of describing specific embodiments only and are not intended to limit this application.

[0034] Fire protection certification exams are administered to firefighters to assess their fire safety knowledge, skills, and emergency response capabilities. Passing these exams demonstrates the necessary competence and qualifications for firefighting. These exams typically assess both theoretical knowledge and practical skills, covering knowledge and skills in fire prevention and control, emergency evacuation, and firefighting. Passing these exams directly impacts a firefighter's competence and responsibility in fire safety work, and therefore holds significant practical significance.

[0035] During the exam, the possibility of various anomalies and security issues must be fully considered, as they could seriously threaten the fairness and security of the exam. For example, cheating may occur through plagiarism or the use of communication devices, which directly affects the fairness of the exam and harms the rights of legitimate candidates. For example, monitoring equipment or systems may not function properly due to technical failures, making it impossible to monitor the exam process in real time, which makes it more difficult to detect anomalies during the exam.

[0036] Furthermore, there may be security vulnerabilities. If hackers attack, exam information may be leaked or tampered with, compromising the security and reliability of the exam. Identity verification issues are also a potential risk. If authentication is incorrect or misused, the accuracy of the candidate's identity may be compromised, affecting the authenticity of the exam results.

[0037] Therefore, in order to ensure the smooth progress and security of the examination process, it is urgent to establish a real-time monitoring system to solve at least one of the above technical problems.

[0038] The embodiment of the present application provides an intelligent identity recognition analysis and security protection system for fire protection examinations. The system includes an access subject and an application trust center, with an encrypted channel set between the access subject and the application trust center.

[0039] The access subject is used to collect real-time surveillance image data of the monitored object; input the real-time surveillance image data into the identity encryption model to generate a real-time login request adapted for the encrypted channel. The real-time login request includes at least: multiply encrypted user identity information, authentication credentials, and behavioral image information; and transmits the real-time login request to the application trust center via the encrypted channel. Thus, by collecting and analyzing the real-time surveillance image data of the monitored object in real time and applying the identity encryption and decryption models to process the identity information, intelligent identification and verification of the monitored object's identity can be achieved, ensuring the accuracy and real-time nature of the monitored object's identity.

[0040] The application trust center is configured to intelligently analyze received real-time login requests using an identity decryption model to obtain real-time identity information of the monitored subject. The real-time identity information includes at least user identity information, authentication credentials, and behavioral image information obtained through multiple decryption processes. Monitoring behavior analysis is performed on the real-time identity information to obtain the exam behavior of the monitored subject. The application trust center then intelligently analyzes the real-time login requests using the identity decryption model to obtain the identity and exam behavior information of the monitored subject. This analysis can help identify abnormal behavior or cheating, ensuring the fairness and accuracy of the exam. A security risk assessment is then performed on the real-time identity information using a security protection model to obtain a security protection assessment associated with the monitored subject. By applying the security protection model to the security risk assessment of the real-time identity information, potential security threats and risks can be promptly identified, helping to provide effective security protection measures for the monitored subject and ensuring the security of the exam process. The corresponding exam behavior and security protection assessment are then labeled for the monitored subject to enable real-time monitoring of the monitored subject. This enables timely detection and recording of important events, providing a basis for subsequent processing and decision-making. Figure 2 This is a schematic diagram of the login interface of the fire protection appraisal test in an embodiment of the present application. Users can enter their account number and password in the login interface to log in.

[0041] Used in the fire identification examination intelligent identity recognition analysis and safety protection system, it performs identity recognition, behavior analysis, and safety protection assessment on real-time monitored subjects during the fire identification examination, helping to improve the security, fairness, and accuracy of the examination process, enhance exam proctoring efficiency, and enhance exam room security. The system can monitor abnormal situations in real time during the examination and promptly identify and take necessary measures to ensure that the fairness and security of the examination are not threatened.

[0042] The intelligent identification analysis and safety protection scheme for fire identification examination provided in the embodiment of the present application can also be executed by an electronic device, which can be a server, a server cluster, or a cloud server. The electronic device can also be a terminal device such as a mobile phone, a computer, a tablet computer, a wearable device, or a special device (such as a special terminal device with an intelligent identification analysis and safety protection system for fire identification examination). These electronic devices can also be equipped with the chips introduced in the above embodiments. Alternatively, these electronic devices can also be installed with a service program for executing the intelligent identification analysis and safety protection scheme for fire identification examination.

[0043] Figure 1A schematic diagram of an intelligent identification analysis and safety protection system for fire protection examination provided in an embodiment of the present application is shown as follows: Figure 1 As shown, the system includes an access subject and an application trust center.

[0044] It is worth noting that an encrypted channel is set up between the access subject and the application trust center.

[0045] In the embodiments of the present application, an encrypted channel refers to a channel that uses encryption technology to encrypt and protect data during the communication process to ensure the security and confidentiality of data transmission. Setting up an encrypted channel between the access subject and the application trust center can effectively prevent data from being stolen, tampered with, or monitored by unauthorized third parties during transmission, ensuring the security of information transmission between the communicating parties. An encrypted channel generally includes the following important components: an encryption algorithm, which is used by the encryption channel to encrypt data. Common encryption algorithms include symmetric encryption algorithms (such as AES) and asymmetric encryption algorithms (such as RSA). Among them, symmetric encryption algorithms are used to protect data during transmission with high encryption speed requirements, and asymmetric encryption algorithms are used in the key negotiation phase before data transmission. A digital certificate is used to prove the identity of the communicating parties and ensure the integrity of the data. A digital certificate generally contains information such as a public key and a digital signature, and can be issued and verified by a certificate authority (CA). A key exchange protocol is used to securely negotiate session keys between the communicating parties. Common key exchange protocols include the Diffie-Hellman key exchange protocol, which ensures that the communicating parties can securely generate a shared key for use in the encrypted communication process. The data integrity protection module,the encrypted channel must also ensure that the transmitted data is not tampered with during,the transmission process. Usually, a message authentication code (MAC) or digital signature,is used to protect the data integrity.

[0046] By setting up an encrypted channel, the communication between the access subject and the application trust center can be fully protected, ensuring that data will not be stolen or tampered with during transmission, thereby ensuring the security and privacy of data transmission.

[0047] It is understood that in the embodiments of the present application, further optionally, deep alternating learning technology is used to implement intelligent encryption protocol selection. That is, the most appropriate encryption algorithm and key exchange protocol are dynamically selected based on the environment and needs of the communicating parties to improve communication efficiency and security.

[0048] Deep alternating learning technology is a novel approach for intelligent encryption protocol selection. It dynamically selects the most appropriate encryption algorithm and key exchange protocol based on the communication environment and requirements of both parties, thereby improving communication efficiency and security. The specific steps are as follows: First, the system collects and preprocesses the communication environment and communication requirements of both parties, including network conditions, communication equipment information, and security policy requirements. Based on this collected data, a deep learning model is designed that learns the optimal encryption protocol selection strategy for both parties from this environment and requirements. Deep alternating learning is an iterative optimization strategy. The system first initializes the encryption algorithm and key exchange protocol using an alternating learning algorithm. Then, in each iteration, the system selects the optimal encryption protocol based on the current environment and requirements, and updates the deep learning model parameters based on the selection results. The system uses the collected data to train and optimize the deep learning model, continuously improving the model's understanding of the communication environment and requirements, thereby achieving more accurate encryption protocol selection. After training is complete, the system applies the learned encryption protocol selection strategy to actual communications. During communication, the system dynamically adjusts the encryption protocol based on real-time environmental changes and communication requirements to ensure communication security and efficiency. Through deep alternating learning technology, the system can continuously learn from the communication process and optimize the encryption protocol selection strategy, achieving smarter and more adaptable encrypted communication, thereby improving communication efficiency and security.

[0049] Multi-layered encryption protection employs multiple layers of encryption, including encryption of transmitted data, data integrity verification, and identity authentication. This creates a more complex and robust encryption channel structure, enhancing data transmission security. See below for an introduction to multiple encryption processes and related encryption and decryption models; this will not be discussed here.

[0050] The following describes the module functions in the system with specific examples.

[0051] The access subject is used to collect real-time monitoring image data of the real-time monitoring object; input the real-time monitoring image into the identity encryption model to generate a real-time login request adapted to the encryption channel; and send the real-time login request to the application trust center through the encryption channel.

[0052] In an embodiment of the present application, the real-time login request adapted to the encrypted channel includes at least: user identity information, authentication credentials, and behavioral image information that have been subjected to multiple encryption processes.

[0053] In applications that implement end-to-end encryption, user identity information is the information provided by users when logging in to verify their identity, such as username, mobile phone number, email address, etc. To protect the security of user identity information, this information is usually encrypted at multiple levels, including symmetric and asymmetric encryption methods, to ensure that the information is not stolen or tampered with during transmission. Authentication credentials are the passwords, PIN codes, fingerprint information, etc. required for user login, which are used to verify the user's identity. In applications that implement end-to-end encryption, authentication credentials are also encrypted to ensure the security of the credentials during transmission. Behavioral image information refers to the behavioral characteristics of users when logging in, such as keystroke speed and mouse movement trajectory, which are used to assist in verifying the user's identity. This information may also be encrypted to ensure that the user's behavioral information is not maliciously stolen or tampered with.

[0054] In real-time login requests, this information is multiply encrypted and transmitted through an end-to-end encrypted channel, ensuring the user's login process is effectively encrypted and protected during transmission, safeguarding the security of the user's identity information and authentication credentials. This combined use of multiple encryption and end-to-end encryption effectively prevents data leakage and unauthorized access, improving application security and the reliability of user authentication.

[0055] For example, consider a smart home security monitoring system that includes cameras for real-time monitoring of the home environment. The access subject here refers to the camera in the smart home security monitoring system, which collects real-time monitoring image data of the monitored object. In this example, the identity encryption model refers to a model used to identify and encrypt user identity information. This model may include technologies such as facial recognition and voiceprint recognition to confirm the user's identity. Once the identity encryption model confirms the user's identity, the system generates a real-time login request adapted for an encrypted channel. This request includes multi-layered encrypted user identity information, authentication credentials, and behavioral image information. The system sends the real-time login request to the application trust center via an encrypted channel. This encrypted channel can use security protocols such as SSL / TLS to protect data transmission security. The application trust center, as the authentication center of the smart home security monitoring system, is responsible for verifying the user's identity and processing the login request. Upon receiving the real-time login request, the application trust center decrypts it and verifies the user's identity.

[0056] In summary, by inputting real-time monitoring image data into the identity encryption model and sending real-time login requests to the application trust center through an encrypted channel, the communication process of the monitoring system can be ensured to be end-to-end encrypted, thereby effectively protecting the privacy and security of users.

[0057] The application trust center is used to use the identity decryption model to perform intelligent analysis on the received real-time login requests to obtain the real-time identity information of the real-time monitored object; perform monitoring behavior analysis on the real-time identity information to obtain the examination behavior of the real-time monitored object; use the security protection model to perform security risk assessment on the real-time identity information to obtain the security protection assessment associated with the real-time monitored object; mark the corresponding examination behavior and security protection assessment for the real-time monitored object to complete real-time monitoring of the real-time monitored object.

[0058] In the embodiment of the present application, the real-time identity information includes at least: user identity information, authentication credentials, and behavioral image information obtained through multiple decryption processes.

[0059] In this embodiment, real-time identity information refers to the user identity information, authentication credentials, and behavioral image information obtained after multiple decryption processes on the data contained in the real-time login request of the real-time monitored object. Similar to the above, the user identity information, the user identity information obtained after the decryption process refers to the information contained in the login request for identifying and verifying the user's identity, such as the user's name, ID number, registered mobile phone number, etc. After decryption, this information can be used to identify the true identity of the current monitored object. The authentication credentials, the authentication credentials after decryption are the credentials used to verify the user's identity, such as the user's password, PIN code, fingerprint information, etc. Through the authentication credentials, the system can confirm whether the identity of the current monitored object is legal. Behavioral image information, behavioral image information refers to the behavioral feature data provided by the monitored object in the login request, such as keystroke speed, mouse movement trajectory, etc. After decryption, this information can be used to analyze the behavioral patterns and habits of the monitored object.

[0060] These real-time identity information is extracted from real-time login requests through decryption processing. After intelligent analysis, monitoring behavior analysis and security risk assessment, the real-time identity information, test behavior and security protection assessment of the real-time monitoring object can be obtained. By marking the corresponding behavior and security assessment for the monitoring object, the application trust center can monitor the monitoring object in real time to ensure the security and reliability of the system. The comprehensive analysis and evaluation of this information can help the application trust center to promptly detect abnormal behavior and security risks and take appropriate measures to deal with them. Figure 3 As shown, this is a schematic diagram of the test answering interface of the fire protection appraisal test in an embodiment of the present application. The behavior of the students will be monitored during the answering process in the test answering interface.

[0061] It's important to understand that real-time identity information is real-time because it's extracted and decrypted from the real-time login requests of the monitored subjects. This means that every time a monitored subject logs in, the system obtains the latest identity information, authentication credentials, and behavioral profile information. Through intelligent analysis, monitoring behavior analysis, and security risk assessment, the system updates and assesses the subject's identity, behavior, and security status in real time. Therefore, the real-time nature of real-time identity information lies in the fact that it is dynamically acquired and updated based on the latest behavior and status of the monitored subject.

[0062] The technical effect of this system can be specifically illustrated through a practical scenario:

[0063] Imagine a fire training center using such a system to monitor the behavior and safety of trainees at a fire certification exam venue. When trainees enter the exam venue, surveillance cameras capture their actions and generate real-time login requests. These requests are sent to the application trust center via an encrypted channel.

[0064] First, the application trust center uses an identity decryption model to intelligently analyze incoming real-time login requests, extracting the student's real-time identity information, such as name and student ID number. This information is updated in real time, reflecting the student's dynamic behavior and location within the examination venue. Furthermore, by monitoring and analyzing real-time identity information, the system can understand student behavior patterns within the examination venue, such as the time they entered the exam room and the duration of their stay. If the system detects abnormal behavior, such as a student attempting to cheat during the exam, it will issue an alarm and log the incident.

[0065] It is understandable that abnormal behavior and cheating behavior are detected by an exemplary method listed below. The system first collects the identity information of the students in real time through the identity decryption model, including name, student number, entry time, etc. This information will be used to build a behavioral baseline model of the students. The system monitors the behavior of the students, including the time of entering and leaving the examination room, the time spent in the examination room, movement trajectory, etc. By comparing with the normal behavior pattern, the system can establish the normal behavior pattern of the students. The system will detect whether the student's behavior deviates from the normal pattern. For example, if a student frequently leaves his seat or frequently talks with others during the exam, this may be marked as abnormal behavior. If the system monitors that the student's real-time location deviates from the examination area, or detects that the student stays in a non-permitted area for too long, it may also be marked as abnormal behavior.

[0066] By deeply analyzing student behavior patterns, the system can identify typical cheating behaviors, such as frequent phone use, frequent interactions with other students, and frequent leaving their seats. These behaviors are typically flagged as potential cheating. The system then identifies anomalous behavior based on known cheating patterns (for example, frequent phone use during the exam period). If it detects behavior matching these patterns, it issues an alert. The system can also cross-validate with other data (such as surveillance camera data and environmental data) to determine whether cheating is occurring. For example, if the system detects that a student has visited multiple prohibited areas within a short period of time, this could be a sign of cheating. When the system detects anomalous behavior or cheating, it triggers an alert mechanism, promptly notifying the invigilator or administrator. The system records detailed information on all anomalous behavior and cheating incidents, including time, location, and type of behavior, for subsequent review and action. Through these steps, the Application Trust Center can intelligently analyze real-time login requests, detect anomalous behavior, and determine whether cheating is occurring, thereby ensuring the fairness and security of exams.

[0067] Next, the system uses a security protection model to conduct a security risk assessment on real-time identity information to evaluate the student's risk to exam security. For example, the system can check whether the student's identity information matches the registration information and whether any unauthorized behavior has occurred. Finally, based on the results of intelligent analysis and security assessment, the system will label each student with the corresponding exam behavior and security protection assessment. For example, if the student's behavior complies with exam regulations and the security risk assessment is low, the system will mark it as compliant. Conversely, if the student's behavior is abnormal and poses a security risk, the system will mark it as abnormal behavior and take appropriate security measures. In this way, the application trust center can monitor the student's behavior and safety in real time within the fire certification examination venue, and promptly detect and respond to potential cheating, thereby ensuring the fairness and reliability of the exam.

[0068] In real-time monitoring scenarios, the system effectively extracts, identifies, and protects the identity information of monitored subjects to combat cheating. The system utilizes an identity decryption model to extract student identity information (such as name and student ID number). By comparing this information with registration information, it ensures the student's identity is authentic and registered. Optionally, biometric technologies such as facial recognition and fingerprint recognition can be combined to further confirm the student's identity and prevent identity theft. The system tracks student dynamic behavior within the examination room in real time, including time of entry and movement trajectory, to ensure information consistency and identity authenticity. The system establishes baseline data on normal behavior patterns, such as student behavior patterns during the exam and common movement patterns within the examination room. By monitoring student behavior in real time, such as frequency of leaving their seat, frequent interactions with others, and use of unauthorized devices, the system detects any anomalies that deviate from normal behavior patterns. By analyzing student behavior patterns, the system identifies potential cheating indicators, such as unusual behavior patterns and similarities to known cheating behaviors. Combining identity information and behavior patterns, the system assesses student security risks, such as whether there has been unauthorized access or high-risk behavior. Based on real-time monitoring and security risk assessments, the system determines whether student behavior complies with exam regulations. If the behavior is normal and the risk assessment is low, it is marked as compliant. If the system detects abnormal behavior with a higher security risk, it is marked as abnormal, records the specific anomaly, and triggers an alert. For students marked as abnormal, the system automatically implements appropriate security measures, such as notifying the proctor or temporarily prohibiting access to certain areas. The system provides real-time feedback to help proctors quickly address anomalies, such as assigning a dedicated person to check on students with suspicious behavior. The system uses an identity encryption model to protect student personal information and ensure data security during transmission and storage. All storage of identity information and monitoring data utilizes high-standard security measures to prevent data leakage and unauthorized access. The system complies with relevant privacy protection laws and regulations, such as the GDPR and China's Personal Information Protection Law (PIPL), to ensure the legal and compliant use of personal information. During processing and analysis, data is anonymized whenever possible to protect student privacy.

[0069] Through these measures, the Application Trust Center can effectively extract, identify and protect students' identity information in real-time monitoring scenarios, and effectively prevent cheating in exams through intelligent analysis and security assessment, thereby ensuring the fairness and reliability of the exams.

[0070] As an optional embodiment, it is assumed that the identity encryption model includes at least a feature extraction layer, an identity recognition layer, an identity authentication layer, a behavior image construction layer, and multiple encryption layers.

[0071] Based on the above structure, when the access subject inputs the real-time monitoring image into the identity encryption model and generates a real-time login request adapted to the encryption channel, it is specifically used to:

[0072] Extracting identity feature information of the real-time monitored object from the real-time monitored image through a feature extraction layer; the identity feature information includes at least: facial feature information, body posture feature information, and action feature information;

[0073] Through the identity recognition layer, the identity feature information is matched to obtain the user identity information corresponding to the real-time monitoring object;

[0074] Through the identity authentication layer, the identity feature information is authenticated and identified to obtain the authentication credential information corresponding to the real-time monitoring object;

[0075] Through the behavior image construction layer, the identity feature information is processed through behavior construction to obtain behavior image information corresponding to the real-time monitoring object;

[0076] The user identity information, authentication credentials, and behavior image information corresponding to the real-time monitoring object are multiply encrypted through multiple encryption layers to construct the real-time login request.

[0077] Specifically, the structure and functions of this identity encryption model can be described as follows:

[0078] The feature extraction layer extracts the subject's identity features from real-time surveillance images, including facial features, body posture features, and motion features. This layer uses image processing and feature extraction algorithms to extract the subject's identity features from surveillance images, providing input data for subsequent identification and authentication.

[0079] The identity recognition layer performs identity matching on the extracted identity features to determine the corresponding user identity information of the monitored object. This layer compares the extracted identity features with known user identity information to determine the identity of the monitored object, such as which employee or trainee it is.

[0080] The identity authentication layer processes the identity information to obtain the corresponding authentication credentials of the monitored object. This layer verifies that the monitored object has valid authentication credentials, such as a card or password, to ensure they are authorized to enter the examination room or office.

[0081] Through the behavior image construction layer, identity feature information is processed through behavior construction to obtain the behavior image information corresponding to the monitored object. The function of this layer is to associate the extracted identity features with the behavior of the monitored object and generate the corresponding behavior image for subsequent behavior analysis and security assessment.

[0082] Multiple encryption layers are used to encrypt the user identity, authentication credentials, and behavioral image information of the monitored subject to construct a real-time login request. This layer protects the privacy of the monitored subject and ensures the security and confidentiality of the real-time login request during transmission.

[0083] In this way, the entire identity encryption model, through the synergy of these layers, can effectively extract, identify and protect the identity information of the monitored object in real-time monitoring scenarios, and generate real-time login requests suitable for encrypted channels to ensure the security and reliability of information transmission.

[0084] In the above embodiment, the identity feature information is subjected to identity matching processing by the identity recognition layer to obtain the user identity information corresponding to the real-time monitoring object, which is expressed as the following formula:

[0085]

[0086] Wherein, F represents the identity feature matrix obtained by extracting the identity feature information through a multi-scale convolutional network, u i represents the i-th identity feature vector in the identity feature information, W represents the learnable weight parameter matrix, W is used to adjust the importance of the identity feature matrix F at different scales, b i represents the learnable bias term, b i Represents the adjustment of the i-th identity feature vector u i Similarity(F,u i ) represents the identity feature matrix F and the i-th identity feature vector u i The similarity between i represents the vector norm. w i That is to say, F·W·u i and the norm of the i-th identity eigenvector ui.

[0087] In the above embodiment, the relationship between corresponding user identity information and similarity can be understood as follows: The identity feature matrix is ​​extracted through a multi-scale convolutional network and contains identity features at different scales. The identity feature vector is a vector of specific identity feature information used for matching. The learnable weight parameter matrix and bias term are used to adjust the importance of the identity feature matrix at different scales and the deviation in the similarity calculation. Similarity is calculated by calculating the similarity between the identity feature matrix and the identity feature vector (for example, using vector inner product or some distance metric). The vector norm represents the length or size of each vector. In the above formula, the calculated similarity is quantified by matching the identity feature matrix and the identity feature vector. This similarity represents the degree of match between the identity feature matrix and the identity feature vector. The corresponding user identity information refers to the specific user identity obtained through comparison during the identity matching process. In other words, the calculated similarity can be used to determine which identity feature vector (i.e., a specific user) best matches the current identity feature matrix. Therefore, the result of the similarity calculation can be used to infer the corresponding user identity information. In summary, similarity itself is not user identity information, but it is used as a basis for inferring or determining the corresponding user identity information. By calculating the similarity, we can find the identity feature vector that best matches the identity feature matrix, thereby obtaining the user's identity information.

[0088] In the above embodiment, the identity feature information is authenticated and identified by the identity authentication layer to obtain the authentication credential information corresponding to the real-time monitoring object, which is expressed as the following formula:

[0089]

[0090] Wherein, F represents the identity feature matrix obtained by extracting the identity feature information through a multi-scale convolutional network, u i represents the i-th identity feature vector in the identity feature information, W represents the learnable weight parameter matrix, W is used to adjust the importance of the identity feature matrix F at different scales, b i represents the learnable bias term, b i Represents the adjustment of the i-th identity feature vector u i The similarity of Ath(F,u i ) represents the identity feature matrix F and the i-th identity feature vector u i The similarity between i represents the vector norm, represents the conversion function that converts similarity into authentication probability.

[0091] This formula compares the feature vector of the real-time monitoring object with the identity feature information of the registered user one by one, calculates the similarity, and converts the similarity into authentication probability through the activation function.

[0092] In this case, there's a conversion process between authentication credentials and authentication probabilities. Authentication credentials may be a set of user characteristics, such as an ID number, fingerprint, or facial features. When the system needs to authenticate a user, it calculates the similarity between the user-provided characteristics and the existing identity characteristics in the system. The authentication probability calculated in the formula is obtained by processing this similarity through an activation function. This authentication probability quantifies the degree to which the system matches the user-provided characteristics with the known characteristics in the system. Therefore, authentication credentials refer to the identity characteristics provided by the user, while the authentication probability refers to the confidence level in the user's identity calculated by the system based on these characteristics. In other words, when the authentication probability reaches a certain threshold, the system deems the submitted authentication credentials valid, and the user is authenticated. In this case, the authentication probability represents a dynamic authentication state, used to determine whether a user's identity has been authenticated. In short, authentication credentials are used to determine whether a user has been authenticated through the calculation of the authentication probability.

[0093] Further optionally, when the identity feature information is processed through the behavior image construction layer to obtain the behavior image information corresponding to the real-time monitoring object, it is specifically used to:

[0094] The posture feature information, movement feature information, expression feature information, and behavioral habit information of the real-time monitoring object are extracted from the identity feature information; the posture feature information, movement feature information, expression feature information, and behavioral habit information of the real-time monitoring object are interwoven and projected from multiple dimensions into a high-dimensional coding space to obtain real-time coding features of the real-time monitoring object; the real-time coding features of the real-time monitoring object are dynamically sampled to obtain initial behavioral image information of the real-time monitoring object; key local areas in the initial behavioral image information of the real-time monitoring object are selected, and image enhancement processing is performed on the key local areas to obtain behavioral image information of the real-time monitoring object.

[0095] This step describes the specific functions of the behavioral image construction layer when further processing identity feature information: First, posture feature information (such as standing or sitting), movement feature information (such as walking or raising a hand), expression feature information (such as smiles and eye contact), and behavioral habit information (such as activity frequency and dwell time) of the real-time monitored subject are extracted from the identity feature information. These features reflect the state and behavior of the monitored subject in the exam or work scenario. Furthermore, the extracted feature information is interwoven from multiple dimensions and projected into a high-dimensional encoding space. This process can be understood as comprehensively considering multiple feature information and representing them as high-dimensional encoding vectors to more comprehensively describe the state and behavior of the real-time monitored subject. Next, the encoding features of the real-time monitored subject are dynamically sampled to capture changes in its real-time state. This method obtains encoding features of the real-time monitored subject at different time points, making the behavioral image information more dynamic and time-series-based. Furthermore, key local regions, such as facial expression regions and hand movement regions, are selected from the initial behavioral image information of the real-time monitored subject and image enhancement processing is performed on these key local regions. This includes enhancing the contrast and clarity of images to highlight key information, such as facial expressions or hand movements, to obtain more accurate and specific behavioral image information.

[0096] The above steps involve extracting and processing multiple features from the identity information of the subject being monitored in real time, and generating behavioral image information in a high-dimensional encoding space. Specifically, information such as posture, movement, expression, and behavioral habits is extracted from the subject's identity information. These features represent the subject's state and behavior in a specific scenario. The extracted features (such as posture and movement) are comprehensively considered and projected into a high-dimensional encoding space. This process fuses features from multiple dimensions into a high-dimensional vector, aiming to more comprehensively describe the subject's state and behavior. This high-dimensional encoding can capture the complex relationships and interactions between features. Dynamic sampling of the encoded features is used to capture the temporal changes in the subject's state. Feature extraction is performed on the subject at different time points, generating a series of encoded features. These time points can be evenly spaced or adjusted based on changes in the subject's behavior. The encoded features at these time points are serialized to form time series data. This allows the system to observe and analyze dynamic changes in the subject's state. Key regions (such as facial expressions and hand movements) are selected from the behavioral image information; these regions are particularly important for behavioral analysis. Furthermore, these key local areas are enhanced, including by improving contrast and clarity, to more accurately extract and analyze detailed information. This helps identify subtle changes and features, thereby improving the accuracy of behavioral image analysis. Sampling techniques (such as sliding windows and time steps) are used to extract encoded features of the object at different time points. This produces a time series, with the encoded features at each time point reflecting the object's state at that point. This time series data can be used to analyze behavioral patterns and detect anomalies. In the high-dimensional encoding space, each encoded vector contains a comprehensive representation of multiple feature information. By analyzing the time series of these vectors, the system can identify trends and patterns in the object's behavior. Enhancement processing makes the information extracted from the image clearer and more prominent, facilitating the detailed analysis of specific behaviors (such as facial expressions and hand movements). In summary, through dynamic sampling, the system can capture the state changes of the real-time monitored object at different time points and encode these changes into high-dimensional vectors, forming complete behavioral image information. By analyzing these encoded vectors, the system can more accurately understand and describe the dynamic behavior of the real-time monitored object.

[0097] Through these steps, the behavioral image construction layer can further transform the identity feature information of the real-time monitored object into behavioral image information that is temporal, dynamic, and information-rich, providing a more accurate and comprehensive data foundation for subsequent behavioral analysis and security assessment.

[0098] Further optionally, the multiple encryption processes include at least one or more of data encryption, tag encryption, time stamp mechanism, data integrity check, anti-tampering processing, and anti-forgery processing. Further optionally, the processing flow parameters in the multiple encryption processes are associated with the encryption channel.

[0099] These multiple encryption processes can be combined with the characteristics of the encryption channel to ensure secure data transmission and integrity. For example, sensitive data in real-time login requests (such as user identity information, authentication credentials, and behavioral image information) is encrypted. The encryption algorithm and key length can be selected based on the requirements of the encryption channel, such as the AES algorithm. Tags are added to the real-time login request to identify the type or source of the data. These tags are encrypted to ensure that only authorized recipients can decrypt and identify them, preventing information leakage or tampering. Timestamps are added to the real-time login request to record the time when the data was generated or transmitted. These timestamps can be signed and encrypted using cryptographic algorithms to ensure data timeliness and integrity, while also preventing replay attacks or tampering with the data's chronological order. Data in the real-time login request is integrity-checked using a hash algorithm to generate a data digest or signature. These digests or signatures can be encrypted using cryptographic algorithms to ensure that the data has not been tampered with or corrupted during transmission. Tamper-evident markers are added to the real-time login request to detect data tampering. Technologies such as digital signatures or message authentication codes (MACs) can be used to authenticate and verify data, ensuring its integrity and credibility during transmission. Authentication mechanisms or digital certificates can be used to authenticate and authorize the sender of real-time login requests. This ensures the source of real-time login requests is legitimate and credible, preventing forgery and malicious attacks.

[0100] The comprehensive use of these multiple encryption processing technologies can effectively protect the data security in real-time login requests and ensure the security and reliability of data transmission on encrypted channels.

[0101] In actual applications, it is further optional to adopt adaptive security strategies to dynamically adjust the security strategies of encrypted channels according to the communication environment and security risks, including the selection of encryption algorithms, adjustment of key lengths, changes in identity authentication methods, etc., to cope with security threats in different scenarios.

[0102] Specifically, adaptive security policy is an intelligent security management method that dynamically adjusts the security policy of encrypted channels by monitoring the communication environment and assessing security risks to address security threats in different scenarios. The specific implementation method is as follows: The system monitors the communication environment, including network status, device information, and access patterns, and assesses security risks, including potential attack types and attack strength, to obtain current security intelligence. Based on the environmental monitoring and risk assessment results, the system dynamically adjusts the security policy of the encrypted channel. This includes selecting the most appropriate encryption algorithm and key length, as well as adjusting the authentication method. Based on the assessment results, the system can select a stronger or lighter encryption algorithm to adapt to the current security threat. For example, in high-risk environments, a more complex symmetric encryption algorithm may be selected, while in low-risk environments, a lighter algorithm may be selected for improved efficiency. Based on the environmental monitoring results, the system can dynamically adjust the key length to ensure protection against current possible attacks. For example, in high-risk environments, the key length can be increased to improve security, while in low-risk environments, the key length can be reduced to improve efficiency. The system can select different identity authentication methods according to actual conditions. For example, two-factor authentication or multi-factor authentication can be used in high-risk environments, while single-factor authentication can be used in low-risk environments to improve convenience.

[0103] Through adaptive security strategies, the system can dynamically adjust the security strategies of encrypted channels according to actual conditions to respond to security threats in different scenarios and ensure the security and efficiency of communications.

[0104] In addition, end-to-end encryption protection can be added to ensure end-to-end encryption protection between the communicating parties, that is, encryption is performed throughout the entire data transmission process to avoid security vulnerabilities in any link of the communication link.

[0105] For example, end-to-end encryption is a key security measure designed to ensure end-to-end encryption between communicating parties, thereby preventing security vulnerabilities at any stage of data transmission. The specific steps for implementing end-to-end encryption are as follows: Before communication begins, the sender encrypts the data to be transmitted. This can be achieved using symmetric or asymmetric encryption algorithms to prevent data theft or tampering during transmission. The encrypted data is transmitted to the receiver via a secure channel. Typically, these channels utilize security protocols (such as SSL / TLS) to protect confidentiality and integrity during data transmission. After receiving the encrypted data, the receiver uses the corresponding key and decryption algorithm to decrypt the data and restore the original data content. This process ensures the confidentiality and integrity of the data during transmission. After the data transmission is completed, the receiver can perform end-to-end verification of the received data to ensure its integrity and authenticity. This can be achieved using methods such as digital signatures or message authentication codes. With end-to-end encryption, both communicating parties can ensure that their communication content is effectively encrypted and protected throughout the transmission process, thereby preventing the possibility of man-in-the-middle attacks or other security threats.

[0106] As an optional embodiment, assuming that the identity decryption model includes at least: a subject identification layer, multiple decryption layers, and an identification layer. Based on the above structure, when the application trust center uses the identity decryption model to intelligently analyze received real-time login requests to obtain real-time identity information of the real-time monitored object, the model is specifically configured to: extract the access subject information in the real-time login request and identify the corresponding access subject type through the subject identification layer; invoke the target decryption module corresponding to the access subject type through the multiple decryption layers to perform multi-layer real-time decryption on the real-time login request to obtain the identity information to be processed; and perform intelligent identification processing on the identity information to be processed through the identification layer to obtain the real-time identity information.

[0107] In the fire identification test scenario, the trust center applies the identity decryption model to intelligently analyze the received real-time login requests to obtain the real-time identity information of the monitored object. The specific application is as follows:

[0108] In the fire protection qualification exam scenario, the subject identification layer first extracts the access subject information from the real-time login request and identifies the corresponding access subject type. For example, the access subject could be a candidate taking the fire protection qualification exam, a fire protection exam administrator, or an invigilator.

[0109] Through multiple decryption layers, each layer calls the corresponding target decryption module based on the access subject's type, performing multi-layer, real-time decryption of the real-time login request. In this scenario, different access subjects may use different authentication methods or credentials. For example, a student may use their ID card information or exam admission ticket number, while an administrator may use a dedicated identity authentication credential.

[0110] The decrypted identity information is then processed intelligently by the recognition layer to obtain real-time identity information. In the fire protection certification exam scenario, the recognition layer will use the decrypted data to identify the candidate's identity information, such as the candidate's name, ID number, and exam session.

[0111] In this way, through extraction and recognition at the subject identification layer, decryption processing at the multiple decryption layers, and intelligent recognition at the identification layer, the trust center can accurately obtain the real-time identity information of candidates taking the fire certification exam. This is crucial for exam management and monitoring, ensuring the security and credibility of the exam.

[0112] As an optional embodiment, a security protection model is a framework or method for assessing and managing system security risks. In this scenario, the security protection model can be used to assess security risks associated with real-time identity information to ensure the security and trustworthiness of the monitored object. Based on this, the application trust center uses the security protection model to perform a security risk assessment on the real-time identity information to obtain a security protection assessment associated with the real-time monitored object. The following steps may be included:

[0113] First, identify security risks associated with real-time identity information, such as identity leakage, identity tampering, and identity forgery. Then, assess the likelihood and impact of these identified risks to determine which are most urgent and severe. Next, implement necessary security controls to mitigate or eliminate these risks, such as strengthening encryption of identity information, implementing access control policies, and enhancing monitoring and auditing. Finally, continuously monitor changes and evolution in security risks and promptly adjust security controls to address new threats and risks.

[0114] In the context of an application trust center, the security protection model can ensure the security and integrity of the identity information of the monitored object during transmission and processing by identifying, evaluating, controlling and monitoring risks of real-time identity information.

[0115] In an embodiment of the present application, a system for intelligent identity recognition, analysis, and security protection for fire safety certification exams is provided. This system performs identity recognition, behavioral analysis, and security protection assessments on real-time monitored subjects during fire safety certification exams, helping to improve the security, fairness, and accuracy of the exam process, as well as the efficiency of exam proctoring and the safety of the exam venue. The system can monitor abnormalities during the exam in real time and promptly identify and take necessary measures to address them, ensuring that the fairness and security of the exam are not threatened.

[0116] In another embodiment of the present application, a method for intelligent identification analysis and security protection of fire identification examination identity is also provided. The method is applied to the access subject in the fire identification examination identity intelligent identification analysis and security protection system, and the system also includes an application trust center, and an encrypted channel is set between the access subject and the application trust center. Figure 4 Said method comprises the following steps:

[0117] 201, collecting real-time monitoring image data of a real-time monitoring object;

[0118] 202. Input the real-time surveillance image into the identity encryption model to generate a real-time login request adapted to the encryption channel; wherein the real-time login request includes at least: user identity information, authentication credentials, and behavioral image information that have been subjected to multiple encryption processes;

[0119] 203. Send the real-time login request to the application trust center through an encrypted channel, so that the application trust center can perform monitoring behavior analysis on the real-time identity information to obtain the examination behavior of the real-time monitoring object, and use a security protection model to perform a security risk assessment on the real-time identity information to obtain a security protection assessment associated with the real-time monitoring object, thereby marking the corresponding examination behavior and security protection assessment for the real-time monitoring object to complete real-time monitoring of the real-time monitoring object.

[0120] It is worth noting that the access subject is also used to perform other functions in the above system, which will not be elaborated here.

[0121] In the embodiments of this application, by accessing the subject, the real-time monitoring objects during the fire protection appraisal examination are identified, their behavior analyzed, and their safety protection assessed. This helps to improve the security, fairness, and accuracy of the examination process, and enhances the efficiency of examination proctoring and the safety of the examination room. The system can monitor abnormal situations during the examination in real time and can promptly identify and take necessary measures to ensure that the fairness and safety of the examination are not threatened.

[0122] In another embodiment of the present application, a method for intelligent identification analysis and security protection of fire identification examination identity is provided. The method is applied to the application trust center of the fire identification examination identity intelligent identification analysis and security protection system, the system also includes an access subject, and an encrypted channel is set between the access subject and the application trust center, see Figure 5 Said method comprises the following steps:

[0123] 301, receiving a real-time login request from a real-time monitored object through an encrypted channel; the real-time login request includes at least: user identity information, authentication credentials, and behavioral image information that have been subjected to multiple encryption processes; the real-time login request is generated through an identity encryption model and is adapted to the encrypted channel.

[0124] 302, using an identity decryption model to intelligently analyze the received real-time login request to obtain real-time identity information of the real-time monitored object; the real-time identity information includes at least: user identity information, authentication credentials, and behavioral image information obtained through multiple decryption processes;

[0125] 303, performing monitoring behavior analysis on the real-time identity information to obtain the examination behavior of the real-time monitored subject;

[0126] 304 , using a security protection model to perform a security risk assessment on the real-time identity information to obtain a security protection assessment associated with the real-time monitoring object;

[0127] 305, marking the corresponding examination behavior and security protection assessment for the real-time monitoring object to complete the real-time monitoring of the real-time monitoring object.

[0128] It is worth noting that the application trust center is also used to perform other functions in the above system, which will not be elaborated here.

[0129] In this embodiment of the application, by using a trust center, real-time monitoring of subjects during the fire protection assessment exam is performed for identity recognition, behavior analysis, and safety assessment, which helps improve the security, fairness, and accuracy of the exam process, as well as the efficiency of exam proctoring and the safety of the exam room. The system can monitor abnormal situations during the exam in real time and can promptly identify and take necessary measures to ensure that the fairness and security of the exam are not threatened.

[0130] In another embodiment of the present application, an electronic device is provided, comprising: a processor, a communication interface, a memory, and a communication bus, wherein, Figure 6 As shown, the processor, communication interface, and memory communicate with each other via a communication bus;

[0131] Memory for storing computer programs;

[0132] The processor is used to implement the intelligent identification analysis and safety protection system for fire protection examination as described in the method embodiment when executing the program stored in the memory.

[0133] The communication bus 1140 mentioned in the electronic device can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The communication bus 1140 can be divided into an address bus, a data bus, a control bus, etc.

[0134] The embodiment of the present application provides a system for constructing a low-energy computing unit for intelligent identity recognition analysis and safety protection in fire identification examinations.

[0135] For ease of representation, Figure 6 Only one thick line is used in the diagram, but this does not mean that there is only one bus or one type of bus.

[0136] The communication interface 1120 is used for communication between the electronic device and other devices.

[0137] The memory 1130 may include a random access memory (RAM) or a non-volatile memory (non-volatile memory), such as at least one disk storage. Alternatively, the memory may be at least one storage device located away from the processor.

[0138] The above-mentioned processor 1110 can be a general-purpose processor, including an artificial intelligence processor, a graphics processing unit (GPU), an artificial intelligence processor card (Machine Learning Unit, MLU), a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0139] Accordingly, an embodiment of the present application further provides a computer-readable storage medium storing a computer program, which, when executed, can implement the steps that can be performed by the electronic device in the above method embodiment.

Claims

1. A fire identification test identity intelligent recognition analysis and safety protection system, characterized by: The system includes an access subject and an application trust center, and an encrypted channel is set between the access subject and the application trust center; wherein, An access subject is configured to collect real-time monitoring image data of a real-time monitoring object; input the real-time monitoring image data into an identity encryption model to generate a real-time login request adapted for an encrypted channel; wherein the real-time login request includes at least: user identity information, authentication credentials, and behavioral image information that have been subjected to multiple encryption processes; and transmit the real-time login request to an application trust center via an encrypted channel; The application trust center is configured to intelligently analyze received real-time login requests using an identity decryption model to obtain real-time identity information of the real-time monitored object; the real-time identity information includes at least user identity information, authentication credentials, and behavioral image information obtained through multiple decryption processes; perform monitoring behavior analysis on the real-time identity information to obtain the examination behavior of the real-time monitored object; perform security risk assessment on the real-time identity information using a security protection model to obtain a security protection assessment associated with the real-time monitored object; and mark the corresponding examination behavior and security protection assessment for the real-time monitored object to complete real-time monitoring of the real-time monitored object. The identity encryption model includes at least a feature extraction layer, an identity recognition layer, an identity authentication layer, a behavior image construction layer, and multiple encryption layers; The access subject inputs the real-time monitoring image into the identity encryption model to generate a real-time login request adapted to the encryption channel, specifically for: Extracting identity feature information of the real-time monitored object from the real-time monitored image through a feature extraction layer; the identity feature information includes at least: facial feature information, body posture feature information, and action feature information; Through the identity recognition layer, the identity feature information is matched to obtain the user identity information corresponding to the real-time monitoring object; Through the identity authentication layer, the identity feature information is authenticated and identified to obtain the authentication credential information corresponding to the real-time monitoring object; Through the behavior image construction layer, the identity feature information is processed through behavior construction to obtain the behavior image information corresponding to the real-time monitoring object. The behavior image information refers to the behavior feature data provided by the monitoring object in the login request, and the behavior feature data includes keystroke speed and mouse movement trajectory; Through multiple encryption layers, the user identity information, authentication credentials, and behavioral image information corresponding to the real-time monitoring object are multiple encrypted to construct the real-time login request; Through the identity recognition layer, the identity feature information is matched to obtain the user identity information corresponding to the real-time monitoring object, which is expressed as the following formula: Wherein, F represents the identity feature matrix obtained by extracting the identity feature information through a multi-scale convolutional network, u i represents the i-th identity feature vector in the identity feature information, W represents the learnable weight parameter matrix, W is used to adjust the importance of the identity feature matrix F at different scales, b i represents the learnable bias term, b i Represents the adjustment of the i-th identity feature vector u i Similarity(F,u i ) represents the identity feature matrix F and the i-th identity feature vector u i The similarity between i represents the vector norm; Through the identity authentication layer, the identity feature information is authenticated and identified to obtain the authentication credential information corresponding to the real-time monitoring object, which is expressed as the following formula: Wherein, F represents the identity feature matrix obtained by extracting the identity feature information through a multi-scale convolutional network, u i represents the i-th identity feature vector in the identity feature information, W represents the learnable weight parameter matrix, W is used to adjust the importance of the identity feature matrix F at different scales, b i represents the learnable bias term, b i Represents the adjustment of the i-th identity feature vector u i The similarity of Ath(F,u i ) represents the identity feature matrix F and the i-th identity feature vector u i The similarity between i represents the vector norm, represents the conversion function that converts similarity into authentication probability.

2. The intelligent identification analysis and safety protection system for fire protection examination according to claim 1 is characterized in that: When the identity feature information is processed through the behavior image construction layer to obtain the behavior image information corresponding to the real-time monitoring object, it is specifically used to: Extracting posture feature information, movement feature information, expression feature information, and behavior habit information of the real-time monitored object from the identity feature information; The posture feature information, movement feature information, expression feature information, and behavior habit information of the real-time monitoring object are interwoven and projected into a high-dimensional coding space from multiple dimensions to obtain the real-time coding features of the real-time monitoring object; Dynamically sampling the real-time coding features of the real-time monitoring object to obtain initial behavior image information of the real-time monitoring object; A key local area in the initial behavior image information of the real-time monitoring object is selected, and image enhancement processing is performed on the key local area to obtain the behavior image information of the real-time monitoring object.

3. The intelligent identification analysis and safety protection system for fire protection examination according to claim 1 is characterized in that: The multiple encryption processing includes at least one or more of data encryption, tag encryption, time stamp mechanism, data integrity check, anti-tampering processing, and anti-counterfeiting processing; The process flow parameters in the multiple encryption process are associated with the encryption channel.

4. The intelligent identification analysis and safety protection system for fire protection examination according to claim 1 is characterized in that: The identity decryption model includes at least: a subject identification layer, multiple decryption layers, and an identification layer; The application trust center uses the identity decryption model to perform intelligent analysis on the received real-time login request to obtain the real-time identity information of the real-time monitored object, specifically for: Extracting the access subject information from the real-time login request through the subject identification layer and identifying the type of the corresponding access subject; Through multiple decryption layers, the target decryption module corresponding to the type of the access subject is called to perform multi-layer real-time decryption on the real-time login request to obtain the identity information to be processed; The identification layer performs intelligent identification processing on the identity information to be processed to obtain the real-time identity information.

5. A method for intelligent identification analysis and safety protection of fire protection examination, characterized in that: The method is applied to an access subject in a fire protection examination identity intelligent recognition analysis and safety protection system, wherein the system further includes an application trust center, and an encrypted channel is provided between the access subject and the application trust center; the method includes: Collect real-time monitoring image data of real-time monitoring objects; Inputting the real-time monitoring image into the identity encryption model to generate a real-time login request adapted to the encryption channel; wherein the real-time login request includes at least: user identity information, authentication credentials, and behavioral image information that have been subjected to multiple encryption processes; The real-time login request is sent to the application trust center through an encrypted channel, so that the application trust center performs monitoring behavior analysis on the real-time identity information to obtain the examination behavior of the real-time monitored object, and uses a security protection model to perform a security risk assessment on the real-time identity information to obtain a security protection assessment associated with the real-time monitored object, thereby marking the corresponding examination behavior and security protection assessment for the real-time monitored object to complete real-time monitoring of the real-time monitored object; The identity encryption model includes at least a feature extraction layer, an identity recognition layer, an identity authentication layer, a behavior image construction layer, and multiple encryption layers; The access subject inputs the real-time monitoring image into the identity encryption model to generate a real-time login request adapted to the encryption channel, specifically for: Extracting identity feature information of the real-time monitored object from the real-time monitored image through a feature extraction layer; the identity feature information includes at least: facial feature information, body posture feature information, and action feature information; Through the identity recognition layer, the identity feature information is matched to obtain the user identity information corresponding to the real-time monitoring object; Through the identity authentication layer, the identity feature information is authenticated and identified to obtain the authentication credential information corresponding to the real-time monitoring object; Through the behavior image construction layer, the identity feature information is processed through behavior construction to obtain the behavior image information corresponding to the real-time monitoring object. The behavior image information refers to the behavior feature data provided by the monitoring object in the login request, and the behavior feature data includes keystroke speed and mouse movement trajectory; Through multiple encryption layers, the user identity information, authentication credentials, and behavioral image information corresponding to the real-time monitoring object are multiple encrypted to construct the real-time login request; Through the identity recognition layer, the identity feature information is matched to obtain the user identity information corresponding to the real-time monitoring object, which is expressed as the following formula: Wherein, F represents the identity feature matrix obtained by extracting the identity feature information through a multi-scale convolutional network, u i represents the i-th identity feature vector in the identity feature information, W represents the learnable weight parameter matrix, W is used to adjust the importance of the identity feature matrix F at different scales, b i represents the learnable bias term, b i Represents the adjustment of the i-th identity feature vector u i Similarity(F,u i ) represents the identity feature matrix F and the i-th identity feature vector u i The similarity between i represents the vector norm; Through the identity authentication layer, the identity feature information is authenticated and identified to obtain the authentication credential information corresponding to the real-time monitoring object, which is expressed as the following formula: Wherein, F represents the identity feature matrix obtained by extracting the identity feature information through a multi-scale convolutional network, u i represents the i-th identity feature vector in the identity feature information, W represents the learnable weight parameter matrix, W is used to adjust the importance of the identity feature matrix F at different scales, b i represents the learnable bias term, b i Represents the adjustment of the i-th identity feature vector u i The similarity of Ath(F,u i ) represents the identity feature matrix F and the i-th identity feature vector u i The similarity between i represents the vector norm, represents the conversion function that converts similarity into authentication probability.

6. A method for intelligent identification analysis and safety protection of fire protection examination, characterized in that: The method is applied to an application trust center in a fire identification test identity intelligent recognition analysis and safety protection system, wherein the system also includes an access subject, and an encrypted channel is provided between the access subject and the application trust center; the method includes: Receiving a real-time login request from a real-time monitored object through an encrypted channel; the real-time login request includes at least: user identity information, authentication credentials, and behavioral image information that have been subjected to multiple encryption processes; the real-time login request generates request data adapted to the encrypted channel through an identity encryption model; An identity decryption model is used to intelligently analyze the received real-time login request to obtain the real-time identity information of the real-time monitored object; the real-time identity information includes at least: user identity information, authentication credentials, and behavioral image information obtained through multiple decryption processes; Performing monitoring behavior analysis on the real-time identity information to obtain the examination behavior of the real-time monitored subject; Using a security protection model, a security risk assessment is performed on the real-time identity information to obtain a security protection assessment associated with the real-time monitoring object; Mark the corresponding examination behaviors and security protection assessments for the real-time monitoring objects to complete the real-time monitoring of the real-time monitoring objects; The identity encryption model includes at least a feature extraction layer, an identity recognition layer, an identity authentication layer, a behavior image construction layer, and multiple encryption layers; The access subject inputs the real-time monitoring image into the identity encryption model to generate a real-time login request adapted to the encryption channel, specifically for: Extracting identity feature information of the real-time monitored object from the real-time monitored image through a feature extraction layer; the identity feature information includes at least: facial feature information, body posture feature information, and action feature information; Through the identity recognition layer, the identity feature information is matched to obtain the user identity information corresponding to the real-time monitoring object; Through the identity authentication layer, the identity feature information is authenticated and identified to obtain the authentication credential information corresponding to the real-time monitoring object; Through the behavior image construction layer, the identity feature information is processed through behavior construction to obtain the behavior image information corresponding to the real-time monitoring object. The behavior image information refers to the behavior feature data provided by the monitoring object in the login request, and the behavior feature data includes keystroke speed and mouse movement trajectory; Through multiple encryption layers, the user identity information, authentication credentials, and behavioral image information corresponding to the real-time monitoring object are multiple encrypted to construct the real-time login request; Through the identity recognition layer, the identity feature information is matched to obtain the user identity information corresponding to the real-time monitoring object, which is expressed as the following formula: Wherein, F represents the identity feature matrix obtained by extracting the identity feature information through a multi-scale convolutional network, u i represents the i-th identity feature vector in the identity feature information, W represents the learnable weight parameter matrix, W is used to adjust the importance of the identity feature matrix F at different scales, b i represents the learnable bias term, b i Represents the adjustment of the i-th identity feature vector u i Similarity(F,u i ) represents the identity feature matrix F and the i-th identity feature vector u i The similarity between i Represents the vector norm.

7. An electronic device, characterized in that: The electronic device comprises: at least one processor, memory, and input-output unit; The memory is used to store computer programs, and the processor is used to call the computer programs stored in the memory to execute the intelligent identity recognition analysis and safety protection system for fire identification examinations as described in any one of claims 1 to 4.

Images