Intelligent examination anti-cheating monitoring system

By building a multi-layer cross-validation mechanism and multi-dimensional data analysis, the blind spots in behavioral cheating detection and proxy exam problems in online exams are solved, full-process monitoring and credit assessment are achieved, and the fairness and reliability of the exams are improved.

CN120782604APending Publication Date: 2025-10-14FENGYE (SHENZHEN) TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202510866756.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-06-26
Publication Date
2025-10-14

AI Technical Summary

Technical Problem

Existing technologies lack a chain of evidence with complementary multimodal data in online exams, making it impossible to comprehensively and accurately detect cheating behavior by examinees. Furthermore, they lack credit assessment of examinees' historical data, making it impossible to effectively prevent cheating and long-term supervision.

Method used

A four-layer cross-validation mechanism of "vision-auditory-thermal-biometrics" is constructed to detect line of sight, body movements and environmental sounds through multi-dimensional data analysis. Combined with iris texture features and historical test data, a multi-source evidence chain is formed to achieve full-process monitoring and credit assessment.

Benefits of technology

It improves the fairness and authority of online exams, accurately identifies cheating behaviors, avoids proxy exams, and enables long-term supervision and credit grading of candidates.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN120782604A_ABST
    Figure CN120782604A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of intelligent examination anti-cheating, and particularly discloses an intelligent examination anti-cheating monitoring system, which comprises a behavior cheating judgment module, an equipment cheating judgment module, a substitute examination cheating judgment module, an examination credit rating evaluation module and a database, according to the method, whether behavior cheating exists or not is judged by obtaining computer screen area pixel boundary coordinates and examinee behavior data, equipment cheating is judged by collecting body surface temperature data and analyzing the temperature rise rate, and cheating for an examinee is analyzed and judged by extracting iris texture features and obtaining gray value deviation. And finally evaluating the credit rating based on the historical examination data. According to the invention, a four-layer cross validation mechanism of visual sense-auditory sense-thermal inductance-biological characteristics is constructed, a single monitoring blind area can be covered by using multi-dimensional data complementation, a multi-source evidence chain is formed to improve the cheating identification accuracy, a comprehensive credit assessment system is established, and long-term supervision of examinees is realized.
Need to check novelty before this filing date? Find Prior Art

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent examination anti-cheating, in particular to an intelligent examination anti-cheating monitoring system. BACKGROUND

[0002] With the rapid development of Internet technology, online examination as a convenient examination form has been widely used in education, professional qualification certification and other fields. However, the openness of the home online examination environment and the limitations of supervision make the examination cheating behavior increasing, which seriously affects the fairness and authority of the examination, therefore, the importance of developing an intelligent examination anti-cheating monitoring system is self-evident.

[0003] The prior art also has the following problems: 1. The four-layer cross-validation mechanism of "vision-audio-thermal-biometric features" is not constructed, and the online examination of examinees is only monitored from the single vision or audio or thermal or biometric feature level, which cannot form a complete evidence chain through multi-modal data complementation, and it is difficult to guarantee the accuracy and comprehensiveness of the whole examination monitoring.

[0004] 2. The existing behavior detection technology can only perform simple action recognition, and cannot perform multi-dimensional comprehensive analysis on the examinee's gaze, body movements and environmental sounds, reducing the comprehensiveness and accuracy of the examinee's examination behavior cheating detection, and easy to appear misjudgment and omission.

[0005] 3. The existing anti-substitute cheating monitoring only confirms the examinee information at the beginning of the examination, and does not continuously detect the biometric features of the examinee during the online examination process, which cannot avoid the occurrence of the examinee's midway substitute behavior.

[0006] 4. The existing technology often only focuses on cheating detection during the examination process, but lacks analysis and credit evaluation of the examinee's historical examination data, which cannot comprehensively evaluate the credit level of the examinee according to the examinee's historical cheating records, and it is difficult to realize the long-term supervision and restraint of the examinee. SUMMARY

[0007] In view of this, in order to solve the problems raised in the background art, an intelligent examination anti-cheating monitoring system is proposed.

[0008] The purpose of the present application can be achieved by the following technical scheme: the present application provides an intelligent examination anti-cheating monitoring system, comprising: a behavior cheating judgment module, which identifies the computer screen area of the target examinee during the home online examination, obtains the pixel boundary coordinates corresponding to the computer screen area, and obtains the behavior data of the target examinee during the home online examination in real time, judges whether the target examinee is a behavior cheating examinee, if yes, immediately feedback, if not, execute the device cheating judgment module.

[0009] The device cheating judgment module collects the body temperature data of the target examinee in the home online examination process, judges whether the target examinee is a device cheating examinee, if yes, immediately feeds back, if not, executes the proxy cheating judgment module.

[0010] The proxy cheating judgment module collects the iris texture features of the target examinee in the home online examination in real time, judges whether the target examinee is a proxy cheating examinee, if yes, immediately feeds back, if not, executes the examination credit level evaluation module.

[0011] The examination credit level evaluation module extracts the historical examination data of the target examinee in the online examination system, evaluates the examination credit level of the target examinee, and feeds back correspondingly, wherein the examination credit level includes first level, second level and third level, first level > second level > third level.

[0012] Compared with the prior art, the embodiments of the present application have at least the following advantages or beneficial effects: (1) The present application constructs a four-layer cross-verification mechanism of "vision-audio-thermal-biometric features", which can use multi-dimensional data to complementarily cover the blind area of single monitoring, such as vision capturing line of sight and body movement, audio recognizing abnormal sound of book turning, thermal positioning device contact causing temperature rise, and biometric feature verifying identity uniqueness, forming a multi-source evidence chain to improve the accuracy of cheating recognition, effectively solving the problems of high missed judgment rate, weak anti-interference ability and insufficient recognition of hidden behaviors in traditional single-layer monitoring, and fully guaranteeing the fairness and authority of online examination.

[0013] (2) The present application can more comprehensively and accurately detect behavior cheating behavior by comprehensively analyzing the line of sight, body movement and environmental sound of the examinee, can accurately judge whether the examinee has abnormal line of sight by coupling analysis of iris center pixel coordinates and stay frame number, can timely find body abnormalities by combining body image and body safety area, and can effectively identify environmental sound abnormalities by matching analysis of environmental sound timbre.

[0014] (3) The present application can accurately judge whether the examinee is a proxy cheating by extracting the gray value of each region of the iris epithelial layer for biometric feature coincidence degree analysis, and can avoid the occurrence of proxy cheating behavior of the examinee by using the uniqueness of iris texture, thereby improving the reliability of proxy cheating detection.

[0015] (4) The present application can comprehensively classify the credit of the examinee according to the cheating record of the examinee by analyzing the historical examination data of the examinee, calculating the examination credit degree, and evaluating the credit level, thereby realizing long-term supervision and constraint of the examinee and improving the cheating cost of the examinee. BRIEF DESCRIPTION OF DRAWINGS

[0016] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings required by the embodiments described below. Obviously, the drawings described below only show some of the embodiments of the present application, and all other drawings obtained by those of ordinary skill in the art without creative labor based on these drawings are within the scope of protection of the present application.

[0017] Figure 1 It is a schematic diagram of the system module structure of the present application.

[0018] Figure 2 It is a flow chart of the examination anti-cheating monitoring of the present application.

[0019] Figure 3 It is a schematic diagram of the behavior cheating judgment step of the present application. DETAILED DESCRIPTION

[0020] The technical solutions of the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor are within the scope of protection of the present application.

[0021] Please refer to Figure 1 and Figure 2 The present application provides an intelligent examination anti-cheating monitoring system, which comprises a behavior cheating judgment module, a device cheating judgment module, a proxy cheating judgment module and an examination credit level evaluation module.

[0022] It should be noted that the present application also includes a database for storing the default body safety area and normal timbre category set of the online examination system, storing the frame interval duration, storing the gray value of each region in the iris epithelium pre-stored by the target examinee when registering, and storing the examination credit interval corresponding to the secondary and tertiary credit levels respectively.

[0023] The behavior cheating judgment module and the device cheating judgment module are connected, the device cheating judgment module and the proxy cheating judgment module are connected, the behavior cheating judgment module, the device cheating judgment module and the proxy cheating judgment module are all connected with the examination credit level evaluation module, and the behavior cheating judgment module, the proxy cheating judgment module and the examination credit level evaluation module are all connected with the database.

[0024] The behavior cheating judgment module identifies the computer screen area of the target examinee during the home online examination, obtains the pixel boundary coordinates corresponding to the computer screen area, and obtains the behavior data of the target examinee in the home online examination process in real time, judges whether the target examinee is a behavior cheating examinee, if yes, immediately feedback, if not, executes the device cheating judgment module.

[0025] It should be noted that the specific way of obtaining the pixel boundary coordinates corresponding to the computer screen region is: detecting the computer screen region edge straight line of the target examinee during the home online examination through Hough transformation, and mapping the three-dimensional screen boundary into the pixel coordinates in the two-dimensional image by combining the perspective transformation matrix, so that the pixel boundary coordinates corresponding to the computer screen region are obtained.

[0026] In the embodiment of the present application, the behavior data includes the corresponding iris center pixel coordinates, iris center pixel coordinate stay frame number, body image and various environmental sound timbres in each capture.

[0027] It should be noted that the video stream of the target examinee is collected by the camera, the iris center is located in real time by using the pupil detection algorithm, the pixel coordinates of each positioning and the frame number (i.e. stay frame number) that the coordinates continuously do not change are recorded, the image containing the body of the target examinee is synchronously collected, and various environmental sound timbres are extracted by means of the microphone and professional audio processing software.

[0028] Please refer to Figure 3 In the embodiment of the present application, the specific process of judging whether the target examinee is a behavior cheating examinee is: extracting the corresponding iris center pixel coordinates, iris center pixel coordinate stay frame number, body image and various environmental sound timbres in each capture from the behavior data of the target examinee during the home online examination.

[0029] Based on the coupling analysis of the corresponding iris center pixel coordinates and iris center pixel coordinate stay frame number in each capture, whether the target examinee has abnormal line of sight during the home online examination is judged.

[0030] In the embodiment of the present application, the specific process of judging whether the target examinee has abnormal line of sight during the home online examination is: comparing the corresponding iris center pixel coordinates with the pixel boundary coordinates corresponding to the computer screen region, and counting the line of sight deviation capture times of the target examinee during the home online examination.

[0031] In the embodiment of the present application, the specific way of counting the line of sight deviation capture times of the target examinee during the home online examination is: if the horizontal coordinate of the corresponding iris center pixel in a capture is less than the horizontal coordinate of the left boundary pixel corresponding to the computer screen region, or greater than the horizontal coordinate of the right boundary pixel corresponding to the computer screen region, or the vertical coordinate of the iris center pixel is less than the vertical coordinate of the upper boundary pixel corresponding to the computer screen region, or greater than the vertical coordinate of the lower boundary pixel corresponding to the computer screen region, then the capture is determined as line of sight deviation capture, so that the line of sight deviation capture times of the target examinee during the home online examination are obtained.

[0032] The frame number of the iris center pixel coordinates corresponding to each time of the line-of-sight deviation capture is extracted from the frame number of the iris center pixel coordinates corresponding to each time of the capture, and multiplied by the frame interval duration stored in the database to obtain the deviation duration corresponding to each time of the line-of-sight deviation capture.

[0033] Based on the number of times of line-of-sight deviation capture of the target examinee during the home online examination and the deviation duration corresponding to each time of the line-of-sight deviation capture, line-of-sight deviation analysis is performed to obtain the line-of-sight deviation degree of the target examinee during the home online examination.

[0034] It should be noted that the specific way of obtaining the line-of-sight deviation degree of the target examinee during the home online examination is to extract the maximum value from the deviation duration corresponding to each time of the line-of-sight deviation capture of the target examinee during the home online examination, and take the maximum value as the deviation duration of the target examinee during the home online examination.

[0035] The number of times of line-of-sight deviation capture and the deviation duration of the target examinee during the home online examination are respectively subtracted from the set reference number of times of line-of-sight deviation capture and the set reference deviation duration, and the two difference values are respectively divided by the set reference number of times of line-of-sight deviation capture and the set reference deviation duration. Finally, the two ratio values are added to obtain the line-of-sight deviation degree of the target examinee during the home online examination.

[0036] The line-of-sight deviation degree of the target examinee during the home online examination is compared with the set reference line-of-sight deviation degree. If the line-of-sight deviation degree of the target examinee during the home online examination is greater than the set reference line-of-sight deviation degree, it is determined that the target examinee has line-of-sight abnormalities during the home online examination, otherwise, it is determined that the target examinee does not have line-of-sight abnormalities during the home online examination.

[0037] Based on the body image corresponding to each time of the capture and the default body safety area of the online examination system stored in the database, it is determined whether the target examinee has body abnormalities during the home online examination.

[0038] It should be noted that the specific process of determining whether the target examinee has body abnormalities during the home online examination is to overlap and compare the body image corresponding to each time of the capture with the default body safety area of the online examination system to obtain the overlapping area of the body image corresponding to each time of the capture.

[0039] The body safety area area is located from the body safety area, and the proportion between the overlapping area of the body image corresponding to each time of the capture and the body safety area area is recorded as the body overlapping area proportion corresponding to each time of the capture.

[0040] The limb overlap area ratio corresponding to each capture is compared with the set reference limb overlap area ratio, if the limb overlap area ratio corresponding to a capture is less than the set reference limb overlap area ratio, it is determined that the target examinee has limb abnormalities during the home online examination, if there is no limb overlap area ratio corresponding to any capture less than the set reference limb overlap area ratio, it is determined that the target examinee does not have limb abnormalities during the home online examination.

[0041] Based on the sound color of each type of environmental sound corresponding to each capture and the set of default normal sound colors of the online examination system in the database, it is determined whether the target examinee has environmental sound abnormalities during the home online examination.

[0042] In the embodiment of the application, the specific way of determining whether the target examinee has environmental sound abnormalities during the home online examination is: matching the sound color of each type of environmental sound corresponding to each capture with the set of default normal sound colors of the online examination system, if the sound color of a type of environmental sound corresponding to a capture is not located in the set of default normal sound colors of the online examination system, the type of environmental sound corresponding to the capture is recorded as an abnormal sound, and the number of times of occurrence of each type of abnormal sound corresponding to each capture is counted.

[0043] The same type of abnormal sound is added to obtain the number of times of occurrence of each type of abnormal sound, and the number of times of occurrence of each type of abnormal sound is compared with the permitted number of times of occurrence of abnormal sound, if the number of times of occurrence of any type of abnormal sound is greater than the permitted number of times of occurrence of abnormal sound, it is directly determined that the target examinee has environmental sound abnormalities during the home online examination, if the number of times of occurrence of any type of abnormal sound is not greater than the permitted number of times of occurrence of abnormal sound, it is determined that the target examinee does not have environmental sound abnormalities during the home online examination.

[0044] If the target examinee has any one of visual line abnormalities, limb abnormalities and environmental sound abnormalities during the home online examination, the target examinee is determined to be a behavior cheating examinee, if there is no any one of visual line abnormalities, limb abnormalities and environmental sound abnormalities, the target examinee is determined not to be a behavior cheating examinee.

[0045] The embodiment of the application can more comprehensively and accurately detect behavior cheating behavior by comprehensively analyzing the visual line, body movement and environmental sound of the examinee, can accurately determine whether the examinee has visual line abnormalities by coupling analysis of the iris center pixel coordinate and the number of frames, can timely find limb abnormalities by combining the body image and the body safety area, and can effectively identify environmental sound abnormalities by matching analysis of the sound color of the environmental sound.

[0046] The device cheating judgment module collects the body surface temperature data of the target examinee in the home online examination process, judges whether the target examinee is a device cheating examinee, if yes, immediately feeds back, if not, executes the proxy cheating judgment module.

[0047] In the embodiment of the present application, the specific process of judging whether the target examinee is a device cheating examinee is: extracting the temperature corresponding to each limb region at each monitoring time point from the body surface temperature data of the target examinee in the home online examination process, and extracting the highest temperature and the lowest temperature from them respectively, and extracting the monitoring time points corresponding to the highest temperature and the lowest temperature, and then obtaining the monitoring interval length between the highest temperature and the lowest temperature.

[0048] It should be noted that the camera with infrared thermal sensing function deployed in the examination terminal performs thermal imaging on the body surface of the target examinee, and identifies the limb regions (hand, arm, trunk) in combination with the human body posture estimation model, segments and extracts the temperature values of the thermal imaging data of each limb region, so as to obtain the temperature corresponding to each limb region at each monitoring time point.

[0049] The extreme value difference between the highest temperature and the lowest temperature of each limb region is compared with the monitoring interval length to obtain the temperature rise rate of each limb region.

[0050] The temperature rise rate of each limb region is compared with the set reference temperature rise rate, if the temperature rise rate of a certain limb region is greater than the set reference temperature rise rate, the limb region is recorded as an abnormal limb region, if there is an abnormal limb region in the home online examination process of the target examinee, the target examinee is determined to be a device cheating examinee, if there is no abnormal limb region, the target examinee is determined to be not a device cheating examinee.

[0051] The proxy cheating judgment module collects the iris texture features of the target examinee in the home online examination, judges whether the target examinee is a proxy cheating examinee, if yes, immediately feeds back, if not, executes the examination credit level evaluation module.

[0052] In the embodiment of the present application, the specific process of judging whether the target examinee is a proxy cheating examinee is: extracting the gray value of each region in the iris epithelial layer from the iris texture features of the target examinee in the home online examination.

[0053] It should be noted that the collection method of the gray value of each region in the iris epithelial layer is: collecting the iris image of the target examinee by the camera, improving the image quality by using image preprocessing technology, performing gray processing on the iris epithelial layer region, dividing different sub-regions, and then counting the gray value distribution of each sub-region, so as to obtain the radial stripes on the iris surface and the gray value of each region in the iris epithelial layer.

[0054] The gray value deviation of each region in the iris epithelial layer of the target examinee is obtained by subtracting the gray value of each region in the iris epithelial layer of the target examinee pre-stored in the database when the target examinee registers from the gray value of each region in the iris epithelial layer of the target examinee when the target examinee takes the home online exam, and the maximum value is extracted as the gray value deviation of the iris epithelial layer.

[0055] The gray value deviation of the set permission is subtracted from the gray value deviation of the iris epithelial layer, and the difference is divided by the gray value deviation of the set permission, so as to obtain the biological feature coincidence degree of the target examinee.

[0056] The biological feature coincidence degree of the target examinee is compared with the biological feature coincidence degree of the set reference, if the biological feature coincidence degree of the target examinee is less than the biological feature coincidence degree of the set reference, it is determined that the target examinee is a substitute cheating examinee, otherwise, it is determined that the target examinee is not a substitute cheating examinee.

[0057] The embodiment of the present application can accurately determine whether the examinee is a substitute cheating examinee by extracting the gray value of each region of the iris epithelial layer and performing biological feature coincidence degree analysis, and the uniqueness of the iris texture is used to avoid the occurrence of the substitute behavior of the examinee in the middle, thereby improving the reliability of the substitute cheating detection.

[0058] The examination credit level evaluation module extracts the historical examination data of the target examinee in the online examination system, evaluates the examination credit level of the target examinee, and performs corresponding feedback, wherein the examination credit level includes first, second and third levels, and the first level>second level>third level.

[0059] In the specific embodiment of the present application, the specific process of evaluating the examination credit level of the target examinee is as follows: the historical examination times, behavior cheating times, device cheating times and substitute cheating times are extracted from the historical examination data of the target examinee in the online examination system.

[0060] If the behavior cheating times, device cheating times and substitute cheating times of the target examinee in the online examination system are all 0, it is determined that the examination credit level of the target examinee is first level.

[0061] If the behavior cheating times, device cheating times and substitute cheating times of the target examinee in the online examination system are not all 0, the examination credit degree of the target examinee is calculated, and is compared with the examination credit degree interval corresponding to the second and third credit levels stored in the database, if the examination credit degree of the target examinee is located in the examination credit degree interval corresponding to the second credit level, it is determined that the examination credit level of the target examinee is second level, if the examination credit degree of the target examinee is located in the examination credit degree interval corresponding to the third credit level, it is determined that the examination credit level of the target examinee is third level.

[0062] In the specific embodiment of the present application, the specific way of calculating the test credit of the target examinee is: adding the behavior cheating times, equipment cheating times and proxy cheating times of the target examinee in the online test system to obtain the total cheating times of the target examinee, recording the proportion between the total cheating times and the historical test times as the cheating times proportion, subtracting the set reference cheating times proportion from the cheating times proportion of the target examinee, and then taking the difference value and the set reference cheating times proportion as a ratio to obtain the test credit of the target examinee.

[0063] The embodiment of the present application calculates the test credit by analyzing the historical test data of the examinee, performs credit level evaluation, can comprehensively classify the credit of the examinee according to the cheating record of the examinee, realizes long-term supervision and constraint of the examinee, and improves the cheating cost of the examinee.

[0064] The embodiment of the present application can utilize multi-dimensional data to complementarily cover the blind area of single monitoring by constructing the four-layer cross verification mechanism of "vision-audio-thermal-biometrics", such as vision capturing line of sight and body movement, audio recognizing abnormal sound such as book turning, thermal positioning device contact causing temperature rise, and biometrics verifying identity uniqueness, forming a multi-source evidence chain to improve the cheating recognition accuracy, effectively solving the problems of high missed judgment rate, weak anti-interference ability, insufficient recognition of hidden behaviors and the like in the traditional single layer monitoring, and comprehensively guaranteeing the fairness and authority of online test.

[0065] The above content is only an example and description of the concept of the present application, and those skilled in the art of the present application can make various modifications or supplements or adopt similar ways to replace the described specific embodiments, as long as the modifications or supplements or replacements do not deviate from the concept of the present application or exceed the scope defined by the present application, and should belong to the protection scope of the present application.

Claims

1. An intelligent examination anti-cheating monitoring system, characterized in that: include: The behavioral cheating judgment module identifies the target candidate's computer screen area during the home online exam, obtains the pixel boundary coordinates corresponding to the computer screen area, and obtains the target candidate's behavioral data during the home online exam in real time to determine whether the target candidate is a behavioral cheating candidate. If so, feedback is immediately provided; if not, the device cheating judgment module is executed; The device cheating detection module collects the target candidate's body surface temperature data during the online exam at home and determines whether the target candidate is a device cheating candidate. If so, feedback is immediately provided. If not, the proxy cheating detection module is executed; The cheating detection module collects the iris texture features of the target candidate in real time while taking the online exam at home, and determines whether the target candidate is a cheating candidate. If so, feedback is immediately provided; if not, the exam credit rating assessment module is executed; The examination credit rating assessment module extracts the historical examination data of the target candidates in the online examination system, evaluates the examination credit rating of the target candidates, and provides corresponding feedback. Among them, the examination credit rating includes level one, level two and level three, and level one > level two > level three.

2. The intelligent examination anti-cheating monitoring system according to claim 1, characterized in that: The behavior data includes the iris center pixel coordinates corresponding to each capture, the number of frames where the iris center pixel coordinates stay, body images, and various environmental sound tones.

3. The intelligent examination anti-cheating monitoring system according to claim 2, characterized in that: The specific process of determining whether the target examinee is a cheating examinee is as follows: Extract the iris center pixel coordinates, the number of frames the iris center pixel coordinates stay in, body images, and various environmental sound tones corresponding to each capture from the target examinee's behavioral data during the online home exam; Based on the corresponding iris center pixel coordinates and the number of frames in which the iris center pixel coordinates remain at each capture, a coupled analysis is performed to determine whether the target candidate has any abnormal vision during the online test at home; Based on the limb images corresponding to each capture and the default limb safety zone of the online examination system stored in the database, determine whether the target candidate has limb abnormalities during the online examination at home; Based on the various environmental sound timbres corresponding to each capture and the default normal timbre set of the online examination system stored in the database, it is determined whether the target candidate has any abnormal environmental sound during the online examination at home; If the target candidate has any of the abnormal vision, abnormal limbs and abnormal environmental sounds during the online examination at home, the target candidate will be judged as a behavioral cheating candidate. If none of the abnormal vision, abnormal limbs and abnormal environmental sounds exist, the target candidate will be judged as not a behavioral cheating candidate.

4. The intelligent examination anti-cheating monitoring system according to claim 3, characterized in that: The specific process of determining whether the target examinee has abnormal vision during the online test at home is as follows: Compare the pixel coordinates of the iris center corresponding to each capture with the pixel boundary coordinates of the computer screen area, and count the number of times the target candidate's gaze deviates during the home online test; Extracting the number of frames in which the iris center pixel coordinate stays corresponding to each sight deviation capture from the number of frames in which the iris center pixel coordinate stays corresponding to each capture, and multiplying it by the frame interval duration stored in the database to obtain the deviation duration corresponding to each sight deviation capture; Based on the number of times the target candidate's gaze deviation is captured during the home online test and the deviation duration corresponding to each gaze deviation capture, the gaze deviation degree of the target candidate during the home online test is obtained; The target candidate's line of sight deviation during the home online examination is compared with the set reference line of sight deviation. If the target candidate's line of sight deviation during the home online examination is greater than the set reference line of sight deviation, it is determined that the target candidate has line of sight abnormality during the home online examination; otherwise, it is determined that the target candidate does not have line of sight abnormality during the home online examination.

5. The intelligent examination anti-cheating monitoring system according to claim 4 is characterized in that: The specific method of counting the number of times the target examinee's gaze deviation is captured during the home online examination is as follows: if the horizontal coordinate of the iris center pixel corresponding to a certain capture is smaller than the horizontal coordinate of the left boundary of the pixel corresponding to the computer screen area, or is larger than the horizontal coordinate of the right boundary of the pixel corresponding to the computer screen area, or the vertical coordinate of the iris center pixel is smaller than the vertical coordinate of the upper boundary of the pixel corresponding to the computer screen area, or is larger than the vertical coordinate of the lower boundary of the pixel corresponding to the computer screen area, then the capture is determined to be a gaze deviation capture, thereby obtaining the number of times the target examinee's gaze deviation is captured during the home online examination.

6. The intelligent examination anti-cheating monitoring system according to claim 3, characterized in that: The specific method for determining whether there is any abnormal ambient sound during the online home exam for the target candidate is: Match the various environmental sound timbres corresponding to each capture with the default normal timbre category set of the online examination system. If the timbre of a certain type of environmental sound corresponding to a certain capture is not within the default normal timbre category set of the online examination system, then the environmental sound corresponding to that capture is recorded as an abnormal sound, and the various types of abnormal sounds corresponding to each capture are counted; Add up the abnormal sounds of the same category to get the number of times each type of abnormal sound occurs, and compare the number of times each type of abnormal sound occurs with the number of times abnormal sounds are allowed to occur. If the number of times any type of abnormal sound occurs is greater than the number of times abnormal sounds are allowed to occur, it is directly determined that there is an abnormal environment sound during the target candidate's online home exam. If the number of times any type of abnormal sound occurs is greater than the number of times abnormal sounds are allowed to occur, it is determined that there is no abnormal environment sound during the target candidate's online home exam.

7. The intelligent examination anti-cheating monitoring system according to claim 1, characterized in that: The specific process of determining whether the target examinee is a device-cheating examinee is as follows: Extract the temperature of each limb area at each monitoring time point from the target candidate's body surface temperature data during the online home exam, and extract the maximum temperature and minimum temperature respectively. At the same time, extract the monitoring time points corresponding to the maximum temperature and minimum temperature, and then obtain the monitoring interval duration between the maximum temperature and the minimum temperature; The temperature rise rate of each limb region is obtained by comparing the extreme difference between the highest temperature and the lowest temperature of each limb region with the monitoring interval. The temperature rise rate of each limb area is compared with the set reference temperature rise rate. If the temperature rise rate of a limb area is greater than the set reference temperature rise rate, the limb area is recorded as an abnormal limb area. If the target candidate has an abnormal limb area during the online test at home, the target candidate is judged to be a device cheating candidate. If there is no abnormal limb area, the target candidate is judged not to be a device cheating candidate.

8. The intelligent examination anti-cheating monitoring system according to claim 1, characterized in that: The specific process of determining whether the target candidate is a cheating candidate is as follows: Extract the grayscale values ​​of each area in the iris epithelium from the iris texture features of the target candidates when taking the online test at home; Subtract the grayscale values ​​of each area in the iris epithelium of the target candidate during the online home test from the grayscale values ​​of each area in the iris epithelium pre-stored by the target candidate when registering in the database, and obtain the grayscale value deviation of each area in the iris epithelium, and extract the maximum value as the grayscale value deviation of the iris epithelium; Subtract the grayscale value deviation of the set permission from the grayscale value deviation of the iris epithelium, and then compare the difference with the grayscale value deviation of the set permission, thereby obtaining the biometric coincidence degree of the target candidate; The target candidate's biometric overlap is compared with the set reference biometric overlap. If the target candidate's biometric overlap is less than the set reference biometric overlap, the target candidate is determined to be a cheating candidate; otherwise, the target candidate is determined not to be a cheating candidate.

9. The intelligent examination anti-cheating monitoring system according to claim 1, characterized in that: The specific process of evaluating the test credit level of the target candidate is as follows: Extract the number of historical exams, number of cheating behaviors, number of cheating devices, and number of cheating by proxy from the historical exam data of the target examinees in the online exam system; If the target candidate's cheating behavior, cheating by device, and cheating by proxy in the online examination system are all 0, the target candidate's examination credit level is determined to be level one; If the target candidate's number of behavioral cheating, device cheating, and proxy cheating in the online examination system is not zero, the target candidate's examination credit is calculated and compared with the examination credit intervals corresponding to the second and third credit levels stored in the database. If the target candidate's examination credit is within the examination credit interval corresponding to the second credit level, the target candidate's examination credit level is determined to be second level; if the target candidate's examination credit is within the examination credit interval corresponding to the third credit level, the target candidate's examination credit level is determined to be third level.

10. The intelligent examination anti-cheating monitoring system according to claim 9, characterized in that: The specific method for calculating the test credit of the target candidate is: adding up the number of behavioral cheating, device cheating and proxy cheating of the target candidate in the online test system to obtain the total number of cheating times of the target candidate, recording the ratio between the total number of cheating times and the number of historical tests as the cheating times ratio, subtracting the set reference cheating times ratio from the target candidate's cheating times ratio, and comparing the difference with the set reference cheating times ratio to obtain the test credit of the target candidate.