A method for analyzing cheating in examinations

Abstract

The present disclosure provides a method for analyzing cheating in an exam. The method includes: obtaining multiple frames of images in a video stream, and determining the offset direction of the test paper of the first object in each frame of the multiple frames; judging whether the test paper has an exposure risk based on the test paper coordinates in each frame of the image and the hand coordinates of the first object; if the test paper has an exposure risk, determining the cheating suspect associated with the first object based on the test paper offset direction; determining the angle feature value, eye contact and writing feature value, and test paper result analysis feature value corresponding to the cheating suspect; determining a comprehensive score based on the angle feature value, eye contact and writing feature value, and test paper result analysis feature value, and determining the degree of cheating suspicion of the cheating suspect according to the comprehensive score. The method realizes the use of intelligent video analysis technology, combined with the correlation analysis of the answer sheet results, to improve the efficiency and accuracy of cheating analysis, reduce the burden of manual review, and improve the accuracy of cheating behavior identification.

Description

Technical Field

[0001] The present disclosure relates to the field of computer technology, and in particular to a method for analyzing cheating in an examination. Background Art

[0002] In modern education, the fairness and effectiveness of examinations are regarded as important criteria for measuring the quality of education and learning outcomes. In order to ensure the fairness of examinations, traditional examination invigilation methods have been widely used, supplemented by video surveillance systems.

[0003] At present, the related technical cheating analysis method records the video data in the examination room in real time to provide the invigilator with an intuitive monitoring screen. The invigilator remotely observes the test status of the examinee through video monitoring and promptly detects and handles suspicious behavior.

[0004] However, the exam cheating analysis methods in related technologies have shown limitations, mainly manifested in the inefficiency and error-proneness of manual review of video data, as well as the lack of analytical capabilities to effectively link video content with answer sheet results, making it difficult to accurately determine whether candidates have actually cheated. Summary of the invention

[0005] The present invention provides a method for analyzing cheating in an examination.

[0006] According to a first aspect of the present disclosure, a method for analyzing cheating in an exam is provided, the method comprising: acquiring multiple frames of images in a video stream, and determining a test paper offset direction of a first object in each frame of the multiple frames; judging whether the test paper has an exposure risk based on the test paper coordinates in each frame and the hand coordinates of the first object; if the test paper has an exposure risk, determining a cheating suspect object associated with the first object based on the test paper offset direction; determining angle feature values, eye contact and writing feature values, and test paper result analysis feature values ​​corresponding to the cheating suspect object; determining a comprehensive score based on the angle feature values, eye contact and writing feature values, and test paper result analysis feature values, and determining the degree of cheating suspicion of the cheating suspect object based on the comprehensive score.

[0007] In some embodiments of the present disclosure, obtaining multiple frames of images in a video stream and determining the test paper offset direction of the first object in each frame of the multiple frames includes: performing color segmentation and edge detection on each frame of the image, respectively, extracting the test paper area and the table area in each frame of the image; determining the distance between the test paper vertex coordinates of the test paper area and the edge center coordinates of the table area; determining the test paper offset result based on the distance and a preset safety threshold; and determining the test paper offset direction based on the test paper offset result.

[0008] In some embodiments of the present disclosure, judging whether there is an exposure risk for the test paper based on the test paper coordinates and the hand coordinates of the first object in each frame image includes: determining the overlapping vertex coordinates of the overlapping area between the test paper area and the hand area based on the test paper vertex coordinates of the test paper area in each frame image and the hand vertex coordinates of the hand area of ​​the first object; if the overlapping vertex coordinates meet the preset overlapping conditions, determining the overlapping area of ​​the overlapping area and the test paper area of ​​the test paper area based on the overlapping vertex coordinates and the test paper vertex coordinates; if the ratio between the overlapping area and the test paper area is less than the preset proportion threshold, determining that there is a preliminary exposure risk for the test paper; determining the overlapping ratio based on the hand vertex coordinates and the test paper vertex coordinates; if the overlapping ratio is greater than or equal to the preset overlapping ratio, determining that there is an exposure risk for the test paper.

[0009] In some embodiments of the present disclosure, determining a suspected cheating object associated with a first object based on a test paper offset direction includes: determining a target analysis population corresponding to the test paper offset direction; and determining the suspected cheating object in the target analysis population using the head posture of each target analysis object in the target analysis population.

[0010] In some embodiments of the present disclosure, determining the target analysis population corresponding to the test paper offset direction includes: determining the associated population corresponding to the first object; determining the bird's-eye view angle of each object in the associated population looking at the first object based on the test paper offset direction and the position coordinates of each object in the associated population; and determining the target analysis population in the associated population based on the bird's-eye view angle.

[0011] In some embodiments of the present disclosure, the head posture of each target analysis object in the target analysis population is used to determine the suspected cheating objects in the target analysis population, including: determining the head deviation angle of each target analysis object according to the key point coordinates of each target analysis object; determining the relative angle between the first object and each target analysis object according to the test paper center point coordinates of each target analysis object and the position coordinates of the first object; determining the angle ratio between the head deviation angle and the relative angle of each target analysis object, and determining the suspected cheating objects in the target analysis population based on the angle ratio.

[0012] In some embodiments of the present disclosure, determining the angle feature value corresponding to the suspected cheating object includes: determining the horizontal deflection angle and the vertical pitch angle of the head of the first object, the seat angle between the first object and the suspected cheating object, and the seat pitch angle between the head of the first object and the desktop of the suspected cheating object; determining the horizontal angle score based on the seat angle between the first object and the suspected cheating object and the horizontal deflection angle of the head of the first object; determining the vertical angle score based on the vertical pitch angle of the head of the first object and the seat pitch angle between the head of the first object and the desktop of the suspected cheating object; determining the total angle score based on the horizontal angle score, the vertical angle score and the preset angle weight; determining the average angle score based on the total angle score and the number of frames corresponding to the video stream; determining the angle standard deviation based on the total angle score and the angle average score; comparing the angle standard deviation with a preset angle standard deviation threshold and the angle average score with a preset angle average score threshold, and obtaining the angle feature value based on the comparison result.

[0013] In some embodiments of the present disclosure, determining the eye writing feature value corresponding to the suspected cheater includes: determining the time the suspected cheater stays in a writing area within a preset monitoring time period, where the writing area stay time includes the gaze stay time and the pen tip stay time; determining the ratio between the writing area stay time and the preset monitoring time period to obtain the writing area stay time ratio; comparing the writing area stay time ratio and a preset stay time ratio threshold, and determining the eye writing feature value based on the comparison result.

[0014] In some embodiments of the present disclosure, determining the test paper result analysis characteristic value corresponding to the suspected cheating object includes: obtaining answer data of the first object and the suspected cheating object, the answer data including answer answers and answer question types; based on the answer answers corresponding to each answer question type, determining the question type similarity corresponding to each answer question type between the first object and the suspected cheating object; based on the number of students and the student scores of each student under each answer question type, determining the score variance corresponding to each answer question type; respectively determining the minimum score variance in the score variance corresponding to each answer question type, and using each minimum score variance as the exam difficulty coefficient corresponding to each answer question type; for each answer question type, determining the cheating suspicion score corresponding to each answer question type based on the question type similarity and the exam difficulty coefficient; and obtaining the test paper result analysis characteristic value based on the preset score weight and the cheating suspicion score corresponding to each answer question type.

[0015] In some embodiments of the present disclosure, a comprehensive score is determined based on angle feature values, eye contact and writing feature values, and test paper result analysis feature values, and the degree of cheating suspicion of the suspected cheating object is determined based on the comprehensive score, including: determining a comprehensive score based on angle feature values, eye contact and writing feature values, test paper result analysis feature values, and preset feature weights; if the comprehensive score is greater than or equal to a preset comprehensive score threshold, determining the degree of cheating suspicion of the suspected cheating object to be high cheating suspicion; if the comprehensive score is less than the preset comprehensive score threshold, determining the degree of cheating suspicion of the suspected cheating object to be low cheating suspicion.

[0016] The cheating analysis method provided by the present disclosure obtains multiple frames of images in a video stream and determines the offset direction of the test paper of the first object in each frame of the multiple frames; based on the test paper coordinates in each frame of the image and the hand coordinates of the first object, determines whether the test paper has an exposure risk; if the test paper has an exposure risk, determines the cheating suspect associated with the first object based on the test paper offset direction; determines the angle feature value, eye contact and writing feature value and test paper result analysis feature value corresponding to the cheating suspect; determines the comprehensive score based on the angle feature value, eye contact and writing feature value and test paper result analysis feature value, and determines the degree of cheating suspicion of the cheating suspect according to the comprehensive score. The use of intelligent video analysis technology combined with the correlation analysis of the answer sheet results greatly improves the efficiency and accuracy of cheating analysis, which not only reduces the burden of manual review and improves the accuracy of cheating behavior identification, but also provides objective and sufficient cheating behavior evidence for educational institutions, facilitating subsequent processing and investigation.

[0017] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The accompanying drawings are used to better understand the present solution and do not constitute a limitation of the present disclosure. Figure 1 A flowchart of a method for analyzing cheating in an exam provided by an embodiment of the present disclosure; Figure 2 A flowchart of another method for analyzing cheating in an exam provided by an embodiment of the present disclosure; Figure 3 A flowchart of another method for analyzing cheating in an exam provided by an embodiment of the present disclosure; Figure 4 A flowchart of another method for analyzing cheating in an exam provided by an embodiment of the present disclosure; Figure 5 A flowchart of another method for analyzing cheating in an exam provided by an embodiment of the present disclosure; Figure 6 A flowchart of another method for analyzing cheating in an exam provided by an embodiment of the present disclosure; Figure 7 A flowchart of another method for analyzing cheating in an exam provided by an embodiment of the present disclosure; Figure 8 A schematic diagram of the structure of a test cheating analysis device provided by an embodiment of the present disclosure; Fig. 9 A schematic diagram of the hardware structure of an electronic device provided in an embodiment of the present disclosure. DETAILED DESCRIPTION

[0019] The following is a description of exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0020] In modern education, the fairness and effectiveness of exams are of vital importance. Traditional exam invigilation methods, even with the aid of video surveillance systems, still have many shortcomings, especially in the post-analysis of cheating behavior. The video cheating analysis method of the related art uses the video surveillance system as an auxiliary tool to record and remotely monitor the exam status in real time, so that the invigilator can promptly detect and deal with suspicious behavior, and save the video recording for subsequent review and analysis. However, with the development of technology and the continuous renovation of cheating methods, the main defects of the video cheating analysis method of the related art are: 1. Low efficiency of manual review: The massive amount of video data makes manual review a huge workload, time-consuming and error-prone, making it difficult to effectively detect hidden cheating behaviors.

[0021] 2. Lack of correlation analysis capabilities: Related technologies can usually only view video recordings in isolation, and cannot correlate video content with the test takers’ answers. For example, if a test taker frequently looks in a certain direction in the video, but it is impossible to compare it with the test takers’ answers in that direction, it is difficult to determine whether there is plagiarism.

[0022] 3. Insufficient post-analysis methods: The relevant technologies lack effective tools and methods to conduct comprehensive analysis of videos and answer sheets after the exam to generate objective and comprehensive cheating analysis results. This makes it difficult for educational institutions to effectively trace and deal with cheating.

[0023] Therefore, the video cheating analysis method of the related technology cannot give full play to the role of video surveillance and ensure the fairness of the examination. This makes educational institutions face a credibility crisis and also affects students' learning enthusiasm and fair competition environment.

[0024] In order to solve the problems in the related technology, the cheating analysis method proposed in this disclosure analyzes multiple frames of images in the video stream to accurately determine the deviation direction and exposure risk of the test paper, and then locks in the suspected cheating object. At the same time, combined with the angle feature value, the eye writing feature value and the test paper result analysis feature value, the degree of cheating suspicion is comprehensively evaluated, and the cheating behavior is efficiently and accurately identified. This not only improves the intelligence level of test supervision, but also effectively guarantees the fairness and impartiality of the test.

[0025] The following describes the test cheating analysis method, device, electronic device, storage medium and computer program product of the embodiments of the present disclosure with reference to the accompanying drawings.

[0026] Figure 1 The following is a flow chart of a method for analyzing cheating in an exam provided by an embodiment of the present disclosure. Figure 1 As shown, the method includes: Step 101: Acquire multiple frames of images in a video stream, and determine the test paper offset direction of the first object in each frame of the multiple frames.

[0027] In some embodiments, the present disclosure can capture a video stream through a camera and extract multiple continuous frames of images from it. Specifically, after obtaining multiple frames of images, the present disclosure can use image processing technology (such as color segmentation, edge detection, contour recognition, etc.) to accurately extract the test paper area and the table area in each frame of the image.

[0028] After determining the test paper area and the table area, the present disclosure can also determine the offset direction of the test paper relative to the table by analyzing the coordinates of the test paper area and the table area. Specifically, the relative positions of the center points or specific points of the two sets of coordinates are compared to determine whether the test paper is offset to the left, right, upward or downward from the table.

[0029] Specifically, the present disclosure can calculate the distance between the vertex coordinates of the test paper area and the center coordinates of the edge of the table area, and this distance reflects the deviation between the test paper and the preset position of the table. Then, based on this distance, a preset safety threshold is compared to obtain a displacement result, that is, to determine whether the test paper is offset and the current offset direction. Finally, based on this offset result, it can be determined in which direction (such as upper left, lower right, etc.) the test paper has been offset.

[0030] Step 102: Based on the test paper coordinates and the hand coordinates of the first object in each frame of the image, determine whether there is an exposure risk in the test paper.

[0031] In some embodiments, when processing each frame of an image in a video stream, the present disclosure can also determine whether there is a risk of the test paper being exposed by the hands of the first object (such as the examinee).

[0032] Specifically, the present disclosure can first determine the overlapping area and its overlapping vertex coordinates between the two based on the vertex coordinates of the test paper area and the vertex coordinates of the hand area. If the vertex coordinates of the overlapping area meet the preset overlapping conditions, it means that the test paper and the hand are in contact or close to contact. Next, the present disclosure can calculate the ratio of the area of ​​the overlapping area to the total area of ​​the test paper. If this ratio is lower than the preset ratio threshold, it indicates that although there is overlap, the area is not large. At this time, the present disclosure can determine that there is a preliminary exposure risk for the test paper.

[0033] In order to more accurately assess the risk, the present disclosure can also directly calculate the overlap ratio of the test paper based on the hand vertex coordinates and the test paper vertex coordinates. If the overlap ratio reaches or exceeds the preset exposure threshold, the present disclosure can determine that the test paper does have an exposure risk.

[0034] When it is determined that there is indeed a risk of exposure of the test paper, the present invention can also provide a real-time exposure risk reminder to prompt the invigilator to take corresponding measures to prevent test paper leakage or improper behavior, or can directly provide an intelligent voice reminder to prompt the first subject to adjust his posture to avoid test paper leakage.

[0035] Step 103: If there is a risk of exposure of the test paper, a cheating suspect object associated with the first object is determined based on the deviation direction of the test paper.

[0036] In some embodiments, in a monitoring scenario, after determining that the test paper is at risk of exposure, the present disclosure can identify potential sight intersection areas or paths where the test paper may be peeped at based on the offset direction of the test paper, and then focus on specific groups of people located in this direction or path, i.e., associated groups. The associated groups may include examinees who are relatively close to the first object in the examination room. Due to their geographical location, they may be able to see the test paper content that is partially or completely exposed due to the offset.

[0037] In order to screen out the suspected cheating object from the target analysis population, the present disclosure can first screen out the target analysis population in the associated population, and the target analysis population refers to the population that can see the test paper of the first object. Then observe and analyze the head posture of each target analysis object, such as the direction of their head, the specific direction of their sight, and whether there are frequent head turns and other actions. These behavioral characteristics can reveal the obvious attention or abnormal behavior of the target analysis object to the test paper. In particular, if the head posture of a certain target analysis object obviously points to the direction of the test paper, or their sight stays in the area exposed by the test paper for a long time, then this target analysis object is likely to be regarded as a suspected cheating object, and the suspected cheating object is an object with potential cheating behavior.

[0038] In addition, after the suspected cheater is identified, the present disclosure can also send a suspected cheater notification to the invigilator of the examination room of the current suspected cheater through remote control, so as to quickly remind the invigilator to focus on the suspected cheater so as to conduct more detailed observation and necessary intervention measures. In this way, not only can the efficiency of invigilation be effectively improved, but also the fairness and justice of the examination can be maintained to the greatest extent, ensuring that every candidate can compete in a fair environment and show their true level.

[0039] Step 104: Determine the angle feature value, eye contact and writing feature value, and test result analysis feature value corresponding to the suspected cheater.

[0040] In some embodiments, after determining the suspected cheater, the present disclosure can combine post-analysis or real-time monitoring analysis to determine the suspected cheater's potential cheating behavior in the examination environment. By analyzing multiple feature values, it is determined whether the suspected cheater has cheated. These feature values ​​mainly include angle feature values, eye contact and writing feature values, and test paper result analysis feature values.

[0041] The angle feature value is used to evaluate whether the suspected cheater has abnormal behavior in terms of head and seat angles. The present disclosure can first collect head and seat angle data about the first object (such as a reference student) and the suspected cheater. This includes the horizontal deflection angle of the head of the first object, the vertical pitch angle of the head, the seat angle between the first object and the suspected cheater, and the seat pitch angle between the head and the suspected cheater's desk. Based on these data, the horizontal angle score and the vertical angle score are calculated, and then the total angle score is determined in combination with the preset angle weight. By considering the number of frames of the video stream, the present disclosure can further calculate the angle average score and the angle standard deviation. Finally, the angle standard deviation and the angle average score are compared with the preset threshold value to determine the angle feature value.

[0042] The eye contact and writing feature value is used to evaluate whether the suspected cheater transmits information through eye contact or writing behavior, so as to determine whether there is cheating behavior. The present disclosure can monitor the residence time of the suspected cheater in the writing area within a preset monitoring time period, including the time of eye contact and the time of pen tip contact. By calculating the ratio between the residence time in the writing area and the preset monitoring time period, the residence time ratio in the writing area is obtained. This ratio is compared with the preset residence time ratio threshold to determine the eye contact and writing feature value.

[0043] The test paper result analysis characteristic value is used to associate with the answer test paper, evaluate the abnormal behavior of the cheating suspect in the answer results, and thus determine whether there is cheating behavior. The present disclosure can obtain the answer data of the first object and the cheating suspect, including the answer and the answer type. By comparing the answers of the two on each answer type, the question type similarity is calculated. At the same time, based on the number of students and the scores of each student under each answer type, the score variance corresponding to each answer type is further determined, and the minimum score variance is found as the test difficulty coefficient. Combined with the question type similarity and the test difficulty coefficient, the cheating suspicion score corresponding to each answer type is calculated. Finally, according to the preset score weight, the test paper result analysis characteristic value is obtained.

[0044] The present invention can more accurately judge whether a suspected cheater has cheated by comprehensively analyzing angle feature values, eye contact and writing feature values, and test paper result analysis feature values, thereby providing powerful technical support for test invigilation.

[0045] Step 105: Determine a comprehensive score based on the angle feature value, the eye contact and writing feature value, and the test result analysis feature value, and determine the degree of cheating suspicion of the suspected cheater based on the comprehensive score.

[0046] In some embodiments, after detecting that a student may have cheated (i.e., after determining the suspected cheater), the present disclosure may introduce a confidence scoring module to quantify the cheating risk of each student. This module will be based on the calculation results of multiple feature values, combined with different weights and intervals, and finally derive a comprehensive score to determine the degree of suspected cheating of the student inferred above.

[0047] The present disclosure can set a feature weight for each feature value (angle feature value, eye contact writing feature value, test paper result analysis feature value). These feature weights can be set based on the importance and reliability of each feature value in judging cheating behavior, and the specific values ​​are not limited in the present disclosure. The present disclosure calculates a comprehensive score based on these feature weights and the corresponding feature values. The comprehensive score is a comprehensive indicator that reflects the degree of abnormal behavior of the suspected cheating object in multiple aspects. The higher the comprehensive score, the higher the degree of suspected cheating of the suspected cheating object.

[0048] Specifically, a comprehensive score is determined based on the angle feature values, eye contact and writing feature values, and test paper result analysis feature values, and the degree of cheating suspicion of the suspected cheater is determined based on the comprehensive score, including: determining a comprehensive score based on the angle feature values, eye contact and writing feature values, test paper result analysis feature values, and preset feature weights; if the comprehensive score is greater than or equal to a preset comprehensive score threshold, determining the degree of cheating suspicion of the suspected cheater to be high; if the comprehensive score is less than the preset comprehensive score threshold, determining the degree of cheating suspicion of the suspected cheater to be low.

[0049] In an optional embodiment of the present disclosure, the present disclosure can calculate the comprehensive score by the following formula:

[0050] in, is the angle eigenvalue, Write the feature value for the eyes, Analyze the eigenvalues ​​for the test results, is the feature weight corresponding to the angle eigenvalue, The feature weight corresponding to the eye contact feature value, The feature weights corresponding to the feature values ​​of the test results are analyzed. ≥Preset comprehensive score threshold , indicating that the degree of suspicion of cheating is high, which is a high suspicion of cheating, and special attention should be paid to the suspected cheater; if the comprehensive score is <Preset comprehensive score threshold , indicating that the degree of suspicion of cheating is low, which is low suspicion of cheating, and there is no need to pay special attention to the suspected cheating object.

[0051] At the same time, in addition to determining the degree of cheating suspicion, the present disclosure can also generate a detailed cheating analysis report based on the obtained degree of cheating suspicion. The cheating analysis report can include the various feature values ​​of the cheating suspect (angle feature value, eye contact and writing feature value, test paper result analysis feature value), comprehensive score and the final cheating suspicion degree judgment result. The existence of the analysis report helps the invigilator to quickly understand the situation of the cheating suspect, and can also provide a strong basis for subsequent further investigation or handling.

[0052] It should be noted that the setting of feature weights, the calculation method of the comprehensive score, and the determination of the comprehensive score threshold in the present disclosure can all be adjusted and optimized according to actual conditions.

[0053] In summary, the technical solution provided by the present disclosure obtains multiple frames of images in a video stream and determines the test paper offset direction of the first object in each frame of the multiple frames; based on the test paper coordinates in each frame and the hand coordinates of the first object, determines whether the test paper has an exposure risk; if the test paper has an exposure risk, determines the cheating suspect associated with the first object based on the test paper offset direction; determines the angle feature value, eye contact and writing feature value, and test paper result analysis feature value corresponding to the cheating suspect; determines the comprehensive score based on the angle feature value, eye contact and writing feature value, and test paper result analysis feature value, and determines the degree of cheating suspicion of the cheating suspect according to the comprehensive score. The use of intelligent video analysis technology combined with the correlation analysis of answer sheet results greatly improves the efficiency and accuracy of cheating analysis, which not only reduces the burden of manual review and improves the accuracy of cheating behavior identification, but also provides educational institutions with objective and sufficient evidence of cheating behavior, facilitating subsequent processing and investigation.

[0054] As a possible implementation, Figure 2 The flowchart of another method for analyzing cheating in an examination is shown. Based on the above embodiment, a specific process of acquiring multiple frames of images in a video stream and determining the offset direction of the examination paper of the first object in each frame of the multiple frames of images includes the following steps: Step 201: Perform color segmentation and edge detection on each frame of image, and extract the test paper area and table area in each frame of image.

[0055] In some embodiments, when processing video frames or images, color segmentation and edge detection techniques can be applied respectively to extract the test paper area and the table area. Color segmentation is performed on each frame of the image. First, the image is converted to the HSV color space, and the HSV threshold is set to identify and extract the color area of ​​the test paper and the table. Then, morphological operations are performed to remove noise and fill holes. Edge detection is then performed, using the Canny algorithm to detect image edges, and the Suzuki-Abe algorithm to extract contours, and finally the test paper and table areas are accurately identified and extracted.

[0056] The purpose of color segmentation is to identify and extract the color areas of the test paper and the table from the image. The purpose of edge detection is to extract the edge contours of the test paper and the table from the image.

[0057] In an optional embodiment of the present disclosure, the present disclosure may first convert the color space of the image into the HSV color space. That is, the image is converted from the BGR color space to the HSV color space using the cv2.cvtColor function of OpenCV. The HSV color space is more suitable for color threshold processing because it separates hue, saturation, and value, making color recognition more accurate.

[0058] According to the color of the test paper and the table, set the corresponding HSV threshold (H ∈ [H_min, H_max] S ∈ [S_min, S_max] V ∈ [V_min, V_max]) to accurately identify the color of the test paper and the table. Use OpenCV's cv2.inRange function to create a binary mask based on the set HSV threshold to extract the area within the specified color range, for example, Mask(x, y) = { if H(x, y) ∈ [H_min, H_max] and S(x, y) ∈ [S_min, S_max] and V(x, y) ∈ [V_min, V_max] take 1, otherwise take 0}.

[0059] For the area within the specified color range, morphological operations (such as opening and closing operations) are used to remove noise and fill holes to obtain a cleaner outline, for example, FinalImage = Closing (Opening (I)), where Opening (I) = (I ⊖ B) ⊕ B and Closing (I) = (I ⊕ B) ⊖ B. At the same time, the binary mask is used to perform a bitwise AND operation with the original image to extract the color areas of the test paper and the table.

[0060] After obtaining the color areas of the test paper and the table, you can first use OpenCV's cv2.cvtColor function to convert the color image into a grayscale image. Then use OpenCV's cv2.GaussianBlur function to perform Gaussian blur on the grayscale image to reduce noise. Use OpenCV's cv2.Canny function to perform Canny edge detection. And use the Suzuki-Abe algorithm (implemented by cv2.findContours function in OpenCV) to extract contours from the edge image, so as to accurately extract the test paper area and the table area.

[0061] Among them, before the present disclosure performs color segmentation and edge detection on each frame of the image, the present disclosure needs to obtain multiple frames of images in the video stream according to the video, specifically including: reading the examination room monitoring video information and transmitting it to the background; using ffmpeg to execute the frame extraction command to obtain static pictures per second from the video stream; classifying and marking the extracted pictures, and the pictures contain meta information such as timestamps and student data. After that, use openpose to extract the pixel coordinates in the picture, specifically including: setting the parameter configuration of OpenPose; specifying the model to be called, selecting FACE_MODEL / BODY_25 / HAND_MODEL as the model for facial key point recognition; initializing the OpenPose instance, configuring and starting facial detection; loading the image to be processed; calling OpenPose to process the image, identifying the facial / body key points / hand key points; outputting the detected facial / body key points / hand key points data, and displaying the detection result image for subsequent cheating analysis.

[0062] Step 202: Determine the distance between the test paper vertex coordinates of the test paper area and the edge center coordinates of the table area.

[0063] In some embodiments, the present invention can obtain the test paper vertex coordinates A (x1, y1), B (x2, y2), C (x3, y3), D (x4, y4) of the four corners of the test paper and the edge center coordinates X (x5, y5), Y (x6, y6), Z (x7, y7), T (x8, y8) of the table area.

[0064] Calculate the distance A between point A and the center X of the table:

[0065] Calculate the distance C between point C and the center Y of the bottom edge of the table: Calculate the distance B between point B and the center Z of the left side of the table:

[0066] Calculate the distance D between point D and the center T of the right side of the table:

[0067] Through the above formula, the present disclosure can obtain the distance between each vertex of the test paper and the center of the corresponding edge of the table. These distances can reflect the positional relationship of the test paper on the table.

[0068] Step 203: Determine the test paper offset result based on the distance and the preset safety threshold.

[0069] In some embodiments, the present disclosure can set safety thresholds in four directions: Threshold_A, Threshold_C, Threshold_B and Threshold_D. These safety thresholds can be adjusted according to the actual application scenario and the size ratio of the test paper and the table. For example, they can be set to 5% or 10% of the width of the test paper.

[0070] Specifically, for the vertical offset judgment: If distanceA>Threshold_A and distanceA>distanceC+Offset_Vertical, the test paper is considered to be biased upward, where Offset_Vertical is the allowable vertical offset tolerance.

[0071] If distanceC>Threshold_C and distanceC>distanceA + Offset_Vertical, the test paper is considered to be biased downward.

[0072] For horizontal offset judgment: If distanceB>Threshold_B and distanceB>distanceD + Offset_Horizontal, the test paper is considered to be biased to the left, where Offset_Horizontal is the allowable horizontal offset tolerance.

[0073] If distanceD>Threshold_D and distanceD>distanceB + Offset_Horizontal, the test paper is considered to be biased to the right.

[0074] Through the above-mentioned direction judgment, the present disclosure can determine whether the test paper offset result causes a vertical offset or a horizontal offset, and whether the offset direction of the current test paper is the left side or the right side.

[0075] Step 204: Determine the test paper offset direction based on the test paper offset result.

[0076] In some embodiments, after obtaining the test paper offset result, the present disclosure can combine the vertical and horizontal offset results to comprehensively determine the offset direction of the test paper. For example, if both the "biased upward" and "biased to the left" conditions are met at the same time, the test paper offset direction is considered to be biased to the upper left.

[0077] At the same time, the present disclosure can also determine whether the test paper has continuously deviated. Specifically, the present disclosure can record the timestamp of each detection. Set a time window t1 (for example, 3 seconds). If within the time window, the same direction of deviation is detected for m1 consecutive times (for example, 2 times), it is considered that the test paper continues to deviate in this direction.

[0078] For example, a queue is used to store the most recent m1 offset directions. Each time an offset is detected, the offset direction is added to the queue. If all elements in the queue are the same, it is considered that the test paper has been continuously offset.

[0079] In summary, the present invention performs color segmentation and edge detection on each frame of image, accurately extracts the test paper and table area, calculates the distance between the test paper vertex and the center of the table edge, and judges the test paper offset in combination with a preset safety threshold, and then determines the offset direction. This can effectively monitor the position of the test paper on the desktop and improve the accuracy of cheating analysis.

[0080] As a possible implementation, Figure 3 The flowchart of another method for analyzing cheating in an examination is shown. Based on the above embodiment, the specific process of judging whether the examination paper has an exposure risk based on the examination paper coordinates and the hand coordinates of the first object in each frame of the image includes the following steps: Step 301: Based on the test paper vertex coordinates of the test paper area and the hand vertex coordinates of the hand area of ​​the first object in each frame image, determine the overlapping vertex coordinates of the overlapping area between the test paper area and the hand area.

[0081] In some embodiments, the present disclosure can determine the overlapping parts and overlapping vertex coordinates of the two areas based on the vertex coordinates of the test paper area and the hand vertex coordinates of the hand area in each frame image.

[0082] In an optional embodiment of the present disclosure, it is assumed that the test paper vertex coordinates of the test paper area are [(x1, y1), (x2, y2)], where (x1, y1) is the upper left corner and (x2, y2) is the lower right corner. At the same time, the hand vertex coordinates of the hand area are [(x9, y9), (x10, y10)], where (x9, y9) is the upper left corner and (x10, y10) is the lower right corner.

[0083] In order to determine whether there is a risk of test paper exposure, that is, whether the hands block or touch the test paper, the present disclosure needs to calculate the overlapping area.

[0084] The upper-left coordinates of the overlapping area are composed of the larger x-value and the smaller y-value of the upper-left coordinates of the two areas, that is, overlap_x1 = max(x1, x9) and overlap_y1 = min(y1, y9). Similarly, the lower-right coordinates of the overlapping area are composed of the smaller x-value and the larger y-value of the lower-right coordinates of the two areas, that is, overlap_x2 = min(x2, x10) and overlap_y2 = max(y2, y10).

[0085] Step 302: If the overlapping vertex coordinates meet the preset overlapping conditions, then based on the overlapping vertex coordinates and the paper vertex coordinates, determine the overlapping area of the overlapping area and the paper area of the paper area.

[0086] In some embodiments, when the present disclosure obtains the correct overlapping area coordinates [(overlap_x1, overlap_y1), (overlap_x2, overlap_y2)], the present disclosure can determine whether there is an overlap. That is, calculate the overlapping area and the paper area according to the vertex coordinates of the overlapping area, and further determine whether the preset overlapping conditions are met and calculate the proportion of the overlapping area in the paper area.

[0087] If overlap_x1 < overlap_x2 and overlap_y1 > overlap_y2, it is determined that the overlapping vertex coordinates meet the preset overlapping conditions, that is, the two rectangles overlap, otherwise there is no overlap.

[0088] If there is an overlap, the overlapping area overlap_area can be calculated by the following formula:

[0089] where (overlap_x1, overlap_y1) are the upper-left coordinates of the overlapping area, and (overlap_x2, overlap_y2) are the lower-right coordinates of the overlapping area.

[0090] The paper area can be calculated by the following formula:

[0091] where (x1, y1) are the upper-left coordinates of the paper area, and (x2, y2) are the lower-right coordinates of the paper area.

[0092] Step 303: If the ratio between the overlapping area and the paper area is less than the preset proportion threshold, it is determined that there is a preliminary exposure risk for the paper.

[0093] In some embodiments, the present disclosure can calculate the proportion of the overlapping area to the test paper area to determine whether there is a preliminary exposure risk in the test paper.

[0094] That is, the ratio shadow between the overlapping area and the test paper area can be calculated by the following formula:

[0095] Among them, overlap_area is the overlapping area, and paper_area is the test paper area.

[0096] In the present disclosure, the preset proportion threshold can be adjusted according to actual conditions and is not limited in the embodiments of the present disclosure. The present disclosure takes 30% as an example. If the overlapping area (i.e., the blocked area) shadow<30%, it is judged that the test paper may be at risk of exposure, i.e., there is a preliminary exposure risk.

[0097] Step 304: Determine the overlap ratio based on the hand vertex coordinates and the test paper vertex coordinates.

[0098] In some embodiments, the present disclosure further determines the overlap ratio in order to accurately determine whether there is an exposure risk.

[0099] The disclosure can first calculate the distance between the hand and the edge of the test paper: Assuming that the left edge of the test paper is the exposed boundary, the disclosure can calculate the distance between the left edge of the hand (x9) and the left edge of the test paper (x1), that is, risk_distance = (x9 - x1). It should be noted that here it is assumed that the hand is on the left side of the test paper. If the hand is on the right side of the test paper, the calculation method needs to be adjusted accordingly.

[0100] The overlap ratio indicates the position of the hand relative to the width of the test paper, that is, whether the hand covers part of the test paper. Since the width of the hand can be calculated by x10 - x9.

[0101] Therefore, the overlap ratio can be expressed as risk_proportion = .

[0102] Step 305: If the overlap ratio is greater than or equal to the preset overlap ratio, it is determined that there is an exposure risk in the test paper.

[0103] In some embodiments, the preset overlap ratio is a pre-set value used to determine whether the hand covers a sufficient area of ​​the test paper to constitute an exposure risk. In this disclosure, the overlap threshold is 50% as an example, that is, the hand covers half or more of the width of the test paper.

[0104] If the overlap ratio is greater than or equal to 50%, the test paper is judged to have an exposure risk.

[0105] Among them, in order to more accurately determine whether the test paper may be exposed, the present disclosure can combine the timestamp of the image and set a time window of t2 seconds. Within this time window, if the risk of exposure is detected multiple times, it can be judged that the test paper is indeed exposed. For example, if the overlap ratio is greater than or equal to 50%, an exposure risk event is recorded, and the number of exposure risk events is counted within the set time window. If the number of exposure risk events exceeds a certain threshold (for example, m2 exposure risks are detected within t2 seconds), the test paper is judged to be exposed.

[0106] In summary, the present invention can accurately determine whether there is an exposure risk of the test paper based on the vertex coordinates of the test paper and the hand in each frame image, and can effectively monitor and warn of the exposure of the test paper to ensure the fairness and security of the examination or evaluation process.

[0107] As a possible implementation, Figure 4 The flowchart of another method for analyzing cheating in an examination is shown. Based on the above embodiment, the specific process of determining a suspected cheating object associated with a first object based on the offset direction of the examination paper includes the following steps: Step 401, determining the target analysis population corresponding to the test paper deviation direction.

[0108] In some embodiments, determining the target analysis population corresponding to the test paper deviation direction includes: determining the associated population corresponding to the first object; determining the overlooking angle of each object in the associated population looking at the first object according to the test paper deviation direction and the position coordinates of each object in the associated population; and determining the target analysis population in the associated population according to the overlooking angle. The overlooking angle is intended to simulate the perspective of other people in the examination room when observing the first object, so as to evaluate whether they can easily perceive the deviation of the test paper.

[0109] In an optional embodiment of the present disclosure, in order to more accurately define the target analysis population with cheating risks, the present disclosure can use statistical methods such as K-means clustering algorithm to first determine the associated population close to the first object. This method can automatically divide students into several clusters according to the distribution of seats in the examination room, and the positions of students in each cluster are relatively close. That is, the present disclosure first sets the number of clusters K (i.e., the number of clusters) and randomly selects K initial centroids (i.e., the center points of the clusters). Then, the distance between each student and each centroid is calculated, and the student is assigned to the cluster where the centroid closest to it is located. Then, the new centroid of each cluster is recalculated according to the position of the students in the cluster, and the above allocation process is repeated until a convergence state is reached (i.e., the allocation of clusters and the position of the centroid no longer change). After the clustering results converge, the present disclosure can obtain the cluster where the first object (such as examinee A) is located. Since the students in this cluster are relatively close in position, their overlooking angle when looking at the first object may be smaller, and it is easier to perceive the offset of the test paper. Therefore, we can regard the students in this cluster as the associated population that can see the test paper of the first object.

[0110] Among them, factors such as the selection of K value and the selection of initial centroid in the K-means clustering algorithm may affect the clustering results. In the embodiments of the present disclosure, there is no restriction on the specific values, which are mainly based on the actual situation.

[0111] After obtaining the associated population, the present disclosure can further combine the horizontal angle and the vertical angle to evaluate whether other students in the examination room can see the deviation of the test paper of the first object. That is, the present disclosure can first define a reference vector, and based on the position coordinates of each object in the associated population, calculate the relative vector of each object in the associated population relative to the first object. Then, the horizontal angle between these relative vectors and the reference vector is calculated using the dot product formula and the arc cosine function. In order to evaluate the visibility in the vertical direction, the present disclosure can also calculate the overlooking angle of each object in the associated population when looking at the first object based on the height of the first object (obtained from the photo by geometric calculation) and the actual distance between the first object and each object in the associated population (also obtained by geometric calculation). Finally, according to the set overlooking angle threshold, those objects whose overlooking angle is less than or equal to the preset overlooking threshold can be screened out, and they are regarded as the target analysis population that can see the test paper of the first object. This method combines the calculation of horizontal angle and vertical angle, and can more accurately determine which students are most likely to observe and therefore may be involved in cheating in the case of test paper deviation.

[0112] Specifically, given the position coordinates of each student in the examination room and the position coordinates of the first object, the students are divided into k clusters using the k-means algorithm.

[0113] Suppose there are four students, (a1, b1), (a2, b2), (a3, b3), (a4, b4).

[0114] Randomly select k initial centroids, (a1, b1), (a2, b2).

[0115] Calculate the distance between students and the centroid and assign them to the cluster closest to them. The calculation formula is as follows:

[0116] By using the above formula, the distance from each student to the two centroids is calculated, and each student is assigned to the cluster where the nearest centroid is located.

[0117] By calculating the distance from each student to the centroid, we can get the initial student allocation result as follows: Cluster 1 = (a1, b1), (a2, b2) Cluster 2 = (a3, b3), (a4, b4).

[0118] Calculate the new centroid for each cluster: Cluster 1 new centroid = ( , ); Cluster 2 new centroid = ( , ) Repeat the calculation until convergence, recalculate the distance from each student to the new centroid, and redistribute the students. After multiple iterations, the cluster assignment and centroid no longer change, and convergence is achieved. Final clustering results: Cluster 1 = (a1, b1), (a2, b2); Cluster 2 = (a3, b3), (a4, b4). Get the cluster where the first object is located, which is the associated population.

[0119] Then, assuming that the test paper is offset to the right, determine which subjects in the associated population can see the test paper of the first subject.

[0120] Specifically, define the reference vector of the first object (student A): Assume that the coordinates of student A are (a1, b1), and select the horizontal right direction (i.e. 0 degrees) as the reference vector, defined as V0 = (1, 0).

[0121] Calculate the vector of student B (an object in the associated population) relative to student B: Assuming that the coordinates of student B are (a2, b2), calculate the vector Vab = (a1-a2, b1-b2) pointing from student A to student B. Then use the dot product formula and the arccosine function to calculate the horizontal angle.

[0122] For example, using the dot product formula to calculate the angle between V0 and Vab : .

[0123] Then use the inverse cosine function to find the angle : .

[0124] Then calculate the student's height, knowing the actual height of the table , the height of the table in the photo , the height of the student in the photo (the vertical distance from the surface of the table to the head of the student) .

[0125] Known geometric calculation formula .

[0126] According to the geometric calculation formula, the height of student A can be calculated .

[0127] Similarly, the distance between student A and student B can be obtained through geometric calculation.

[0128] Finally, the distance between student A and student B and the height of student A , we can calculate the angle of view of student B looking at student a ,Right now .

[0129] Step 402: using the head posture of each target analysis object in the target analysis population, determine the suspected cheating object in the target analysis population.

[0130] In some embodiments, using the head posture of each target analysis object in the target analysis population, determining the suspected cheating objects in the target analysis population includes: determining the head offset angle of each target analysis object based on the key point coordinates of each target analysis object; determining the relative angle between the first object and each target analysis object based on the test paper center point coordinates of each target analysis object and the position coordinates of the first object; determining the angle ratio between the head offset angle and the relative angle of each target analysis object, and determining the suspected cheating objects in the target analysis population based on the angle ratio.

[0131] Specifically, the present disclosure can determine whether there is cheating behavior by analyzing the field of view direction of the target object's head posture.

[0132] Determining the head posture of the target analysis object specifically includes: using the OpenPose face model to identify the facial key points of the target analysis object, and these facial key points will be used for subsequent posture estimation. In order to perform head posture estimation, the present disclosure can define a standard human face three-dimensional model.

[0133] Define a standard 3D face model, including the 3D coordinates of feature points such as nose tip (0, 0, 0), left eye corner (-x, y, z), right eye corner (x, y, z), left mouth corner (-x', -y', z'), right mouth corner (x', -y', z') and chin (0, -h, -d) (equivalent to an ideal face model). These coordinates are usually provided by actual measurements or standard templates.

[0134] In actual scenes, the mapping relationship between the coordinates in the two-dimensional image and the three-dimensional world coordinates depends on the camera's intrinsic parameters. The camera's intrinsic parameters include: focal length and , usually based on the actual camera settings or image resolution; the optical center , represents the center point of the image.

[0135] Using the pinhole camera model, define the camera's intrinsic parameter matrix (This matrix is ​​used to project three-dimensional coordinates onto the two-dimensional image plane.): .

[0136] Then, the PnP algorithm is used to calculate the rotation matrix and displacement matrix of the head according to the feature point positions in the two-dimensional image and the three-dimensional model points. Through the solvePnP function of OpenCV, the known 3D model points, the corresponding 2D image points, the camera intrinsic parameter matrix and the distortion coefficient are input to obtain the rotation vector and displacement vector. Among them, PnP (Perspective-n-Point) is an algorithm for estimating the pose of a calibrated camera, given a set of 3D points and their 2D projections in the image. The goal of this algorithm is to calculate the rotation and translation of the camera relative to the world coordinate system, with six degrees of freedom (6 DOF), namely the rotation (roll, pitch and yaw) and three-dimensional translation of the camera.

[0137] The pitch angle, yaw angle and roll angle of the head are obtained by converting the rotation vector to the rotation matrix and the rotation matrix to the quaternion. The specific method is as follows: First, the rotation vector r is converted into a rotation matrix using the Rodrigues formula. The rotation vector is usually represented as a three-dimensional vector, which represents the angle of rotation around a unit vector.

[0138] The Rodrigues formula is as follows: , where: R is the rotation matrix, is the rotation angle; is the cross product matrix of the rotation axis; M is the identity matrix.

[0139] Quaternion is a mathematical representation method that extends complex numbers. It is a four-dimensional number composed of four real numbers. Quaternion is mainly used in computer graphics, robotics and physics, especially in the representation and calculation of 3D rotation. Quaternion has higher stability and efficiency than Euler angles and rotation matrices.

[0140] Convert from rotation matrix R to quaternion: q=(w,x,y,z).

[0141] Conversion of quaternion to rotation vector r: , where w is the scalar part of the quaternion and (x, y, z) is the vector part of the quaternion. Where w, x, y, z are the components of the quaternion, the direction of the rotation vector r is given by the x, y, z components of the quaternion, and the rotation angle is Determined by w.

[0142] In practical applications, such as head motion analysis, the present disclosure can obtain the head orientation through a rotation matrix or quaternion. For example, the head offset (such as up and down, left and right, rotation) can be described by calculating a rotation matrix or quaternion.

[0143] Pitch: corresponds to the rotation around the horizontal axis (for example, the tilt of the head up and down). It can be calculated by the components of the quaternion or the rotation vector.

[0144] Yaw: corresponds to the rotation around the vertical axis (for example, the rotation of the head left and right). It can also be calculated using a rotation vector or quaternion.

[0145] Roll: corresponds to the rotation around the front-to-back axis (for example, the roll of the head). The calculation method is similar to the pitch angle.

[0146] The final values ​​are: Pitch angle: [-90°, 90°] Yaw angle: [-180°, 180°] Roll angle: [-180°, 180°] Afterwards, the present disclosure can determine the relative angle between the target analysis object and the first object according to the distribution of seats in the examination room: Assume that the coordinates of the target analysis object are (a2, b2), and the coordinates of the center point of the test paper of student B (i.e., the first object) are (a1, b1).

[0147] Use the cosine function to calculate the relative angle between two points :

[0148] Analyze the head yaw angle (indicates the left-right rotation of the head) and relative angle of the object according to the target , calculate the angle ratio .

[0149] The present disclosure can set a reasonable preset ratio threshold according to actual conditions to determine whether the target analysis object is likely to be peeking at the test paper.

[0150] If the angle ratio is greater than or equal to the preset ratio threshold (or Yaw and , and the difference is in the same direction and within a reasonable range), the current target analysis object is considered to be suspected of cheating, and the current target analysis object is determined to be a suspected cheating object.

[0151] At the same time, the present disclosure can also detect multiple cheating behaviors within a time window: Combine the timestamp of the image and set a time window of n seconds. Within the time window, check whether the target analysis object is detected to have peeking behavior multiple times (i.e., determine whether the number of times the angle ratio of the target analysis object within the time window is greater than or equal to the preset ratio threshold is greater than or equal to the preset number of times). If it is detected multiple times, it is determined that the target analysis object may be cheating, and the target analysis object is a suspected cheating object.

[0152] In summary, the present invention can effectively screen out those suspected of cheating by accurately identifying the target analysis group pointed by the deviation direction of the test paper and further using the head posture information of these students. This not only improves the pertinence of cheating detection, but also ensures the accuracy and efficiency of detection.

[0153] As a possible implementation, Figure 5 The flowchart of another method for analyzing cheating in an examination is shown in FIG. Based on the above embodiment, the specific process of determining the angle feature value corresponding to the suspected cheating object includes the following steps: Step 501: Determine the horizontal deflection angle and vertical pitch angle of the head of the first subject, the seat angle between the first subject and the suspected cheater, and the seat pitch angle between the head of the first subject and the tabletop of the suspected cheater.

[0154] In some embodiments, the horizontal deflection angle of the head of the first object is determined according to the head offset angle and the position coordinates. and head vertical pitch angle , the seat angle between the first object and the suspected cheating object and the seat pitch angle between the first subject's head and the cheating suspect's desk .

[0155] Step 502: Determine a horizontal angle score based on the seat angle between the first subject and the suspected cheating subject and the horizontal deflection angle of the head of the first subject.

[0156] In some embodiments, the absolute difference of the horizontal deflection angle is calculated based on the seat angle between the first object and the suspected cheating object and the horizontal deflection angle of the head of the first object. .

[0157] Then the horizontal deflection angle score, i.e. the horizontal angle score, is calculated using the following formula:

[0158] Indicates the horizontal angle tolerance, for example 3°; Indicates the maximum horizontal angle difference, for example, 30°, used for normalization.

[0159] Step 503: Determine a vertical angle score based on the vertical pitch angle of the first subject's head and the seat pitch angle between the first subject's head and the tabletop of the suspected cheater.

[0160] In some embodiments, the absolute difference in the vertical pitch angle is calculated based on the vertical pitch angle of the first subject's head and the seat pitch angle between the first subject's head and the table of the suspected cheater. .

[0161] Then the vertical pitch angle score, i.e. the vertical angle score, is calculated using the following formula:

[0162] Indicates the vertical angle tolerance; Indicates the maximum vertical angle difference, used for normalization.

[0163] Step 504: Determine the total angle score based on the horizontal angle score, the vertical angle score and the preset angle weight.

[0164] In some embodiments, the present disclosure may calculate the total angle score of each frame of image based on the horizontal angle score, vertical angle score and preset angle weight of each frame of image by applying the following formula: :

[0165] in, is the horizontal angle score for each frame of image, is the preset angle weight corresponding to the horizontal angle score, is the vertical angle score of each frame image, is the preset angle weight corresponding to the vertical angle score, , you can set the specific value of the weight according to the scene requirements, for example =0.7, =0.3.

[0166] Step 505: Determine an average angle score based on the total angle score and the number of frames corresponding to the video stream.

[0167] In some embodiments, the present disclosure can first determine the corresponding frame number of multiple frames in the video stream, and then use the total angle score of each frame and the frame number to calculate the average angle score of the current multiple frames. , as shown in the following formula.

[0168]

[0169] Where N is the number of frames corresponding to multiple frames in the video stream (i.e. the total number of frames in the video segment of the current time period). For the The time point corresponding to the frame image, is the total angle score corresponding to the i-th frame image.

[0170] Step 506: Determine the angle standard deviation based on the angle total score and the angle average score.

[0171] In some embodiments, after obtaining the angle average score of the current multi-frame images, the present disclosure can use the total angle score of each frame image and the angle average score of the multi-frame images to calculate the angle standard deviation of the current multi-frame images (i.e., the current time period). The specific formula is as follows:

[0172] Where N is the number of frames corresponding to multiple frames in the video stream (i.e. the total number of frames in the video segment of the current time period). For the The time point corresponding to the frame image, is the total angle score corresponding to the i-th frame image, It is the average score of the angle of the current multi-frame image.

[0173] Step 507: respectively compare the angle standard deviation with a preset angle standard deviation threshold and the angle average score with a preset angle average score threshold, and obtain an angle feature value based on the comparison results.

[0174] In some embodiments, after obtaining the angle standard deviation, the present disclosure may determine the angle feature value using the angle standard deviation and the angle average score.

[0175] Specifically, the present disclosure can compare the angle standard deviation with a preset angle standard deviation threshold and the angle average score with a preset angle average score threshold. Greater than or equal to the preset angle average score threshold (e.g. 0.8), and the angle standard deviation Less than or equal to the preset angle standard deviation threshold (e.g. 0.3), it means that the cheating possibility of the current cheating suspect is high in most time slices and the behavior pattern is consistent. At this time, the angle feature value is determined. is 1, otherwise the angle eigenvalue is 0.

[0176] In summary, the present disclosure accurately calculates the head rotation angle of the first object, the seat angle with the suspected cheating object, and the seat pitch angle, so that the system can carefully analyze the visual attention of the suspected cheating object to the first object. Combining the scores of horizontal angles and vertical angles, as well as preset weights, the system generates a total angle score, and further calculates the average angle score and the angle standard deviation, so as to comprehensively evaluate the line of sight dynamics of the suspected cheating object. Finally, the calculated angle standard deviation and average score are compared with the preset thresholds respectively. The system can accurately determine and output the angle feature value that reflects the line of sight characteristics of the first object, providing a strong basis for identifying potential cheating behavior.

[0177] As a possible implementation, Figure 6 The flowchart of another method for analyzing cheating in an exam is shown in FIG. Based on the above embodiment, the specific process of determining the eye contact and writing feature value corresponding to the suspected cheater includes the following steps: Step 601: Determine the time that the suspected cheater stays in the writing area within a preset monitoring time period, where the time that the suspected cheater stays in the writing area includes the time that the eyes stay and the time that the pen tip stays.

[0178] In some embodiments, the time that the suspected cheater stays in the writing area of ​​the test paper refers to the time that the eyes or pen tip of the suspected cheater stays in the writing area of ​​the test paper. When determining the time that the suspected cheater stays in the writing area within a preset monitoring time period, the present disclosure comprehensively considers the time that the eyes stay and the time that the pen tip stays. These two indicators together reflect the activities of the suspected cheater in the writing area. To achieve this goal, the present disclosure can use eye tracking equipment and supporting software to accurately track and record relevant data.

[0179] Step 602: Determine the ratio between the residence time in the writing area and the preset monitoring time period to obtain the residence time ratio in the writing area.

[0180] In some embodiments, the present disclosure defines a writing area dwell time ratio (Writing_ratio), which represents the percentage of the time that the cheating suspect's eyes or pen tip stay in the writing area in the entire monitoring time period. The preset monitoring time period of the present disclosure can be set according to actual needs, and the present disclosure takes 30 minutes as an example.

[0181] Step 603: Compare the dwell time ratio of the writing area with a preset dwell time ratio threshold, and determine the eye writing feature value according to the comparison result.

[0182] In some embodiments, the present disclosure pre-sets a dwell time ratio threshold (T_writing) as a reference standard for determining whether the user has been away from the writing area for a long time.

[0183] The present disclosure compares the calculated dwell time ratio of the writing area with the preset dwell time ratio threshold. According to the comparison result, the eye contact writing feature value is determined. The eye contact writing feature value can intuitively reflect the degree of attention paid by the suspect to the writing area during the monitoring period, thereby providing an important basis for the subsequent judgment of cheating behavior.

[0184] Specifically, if the dwell time ratio (Writing_ratio) of the writing area is greater than or equal to the preset dwell time ratio threshold (T_writing), the eye contact writing feature value is determined. 1 indicates that the suspected cheater spends most of his time writing normally. If the proportion of time spent in the writing area (Writing_ratio) is less than the preset threshold of the proportion of time spent (T_writing), the eye contact writing feature value is determined. A value of 0 indicates that the suspected cheater may have potential cheating behavior, such as not looking at or writing the test paper for a long time.

[0185] Among them, the present disclosure can reasonably set and adjust the residence time ratio threshold according to historical data and experience to improve the accuracy and efficiency of cheating behavior identification. The present disclosure takes 70% as an example.

[0186] In summary, the present disclosure uses precise monitoring technology to accurately calculate the time that a suspected cheater stays in the writing area during a preset monitoring period, which covers the length of time that his eyes and pen tip stay in the area. Furthermore, we compare this residence time with the total monitoring time to obtain the residence time ratio in the writing area, which intuitively reflects the degree of concentration of the suspect on the writing area. Subsequently, we compare the obtained ratio with the preset residence time ratio threshold, and based on the comparison results, we can determine and generate an eye-writing feature value. This feature value serves as an important indicator for judging whether a suspected cheater is focused on writing activities, and provides strong data support for subsequent cheating behavior identification.

[0187] As a possible implementation, Figure 7 The flowchart of another method for analyzing cheating in an examination is shown in FIG. Based on the above embodiment, the specific process of determining the characteristic value of the test paper result analysis corresponding to the suspected cheating object includes the following steps: Step 701: Obtain the answering data of the first object and the object suspected of cheating. The answering data includes the answering answers and the question types of the answers.

[0188] In some embodiments, the present disclosure may collect the answering answers and question types of each question of the object suspected of cheating and the first object. At the same time, the present disclosure may also collect the answering data of other students in the same class, which is mainly used for calculating the option probabilities of multiple-choice questions subsequently.

[0189] After obtaining the answering data, the present disclosure may determine the question types of each question, including but not limited to multiple-choice questions, true or false questions, multiple-select questions, fill-in-the-blank questions, and short-answer questions.

[0190] Among them, after obtaining the answering data, the present disclosure may also perform text preprocessing on text question types (fill-in-the-blank questions and short-answer questions), including converting the answers of fill-in-the-blank questions and short-answer questions to lowercase to eliminate the influence of case differences on text similarity calculation; removing punctuation marks in the answers to make the text cleaner and facilitate subsequent processing; performing word segmentation on the answers to split continuous text into meaningful lexical units; performing synonym or near-synonym replacement on the word-segmented results to further improve the accuracy of text similarity. Optionally, the present disclosure may also remove stop words (such as common but meaningless words like "de", "le", etc.) in the answers to reduce noise.

[0191] In addition, for each multiple-choice question type, the present disclosure counts the number of times each option j is selected in the class. Divide the number of times each option is selected by the total number of students in the class to obtain the probability P(j) of the option being selected.

[0192] Step 702: Based on the answering answers corresponding to each question type, determine the question type similarity corresponding to each question type between the first object and the object suspected of cheating.

[0193] In some embodiments, after obtaining the answering data, the present disclosure may calculate the relative student question type similarity. That is, the present disclosure may obtain the answer sets corresponding to each question type of the first object and the object suspected of cheating, and then determine the similarity coefficient between the first object and the object suspected of cheating under each question type based on the answer set corresponding to each question type of the first object and the answer set corresponding to each question type of the object suspected of cheating. Finally, based on the similarity coefficient, determine the question type similarity corresponding to each question type between the first object and the object suspected of cheating.

[0194] Among them, the similarity coefficient and the question type similarity can be calculated by the following formula:

[0195]

[0196] is the similarity coefficient between the first subject and the suspected cheating subject for each question type; is the question type similarity between the first subject and the suspected cheating subject for each question type; A is the answer set of the first subject; B is the answer set of the suspected cheating subject; |A∩B| represents the size of the intersection of A and B (i.e., the number of answers that are the same for the two subjects), and |A∪B| represents the size of the union of A and B (i.e., the number of answers filled in by at least one subject).

[0197] In an optional embodiment of the present disclosure, taking the multiple-choice question type as an example, the present disclosure can obtain the option ranking of a certain multiple-choice question type for an object (assuming it is student A) and a suspected cheating object (assuming it is student B) (for multiple-choice questions, the set of selected options is required to be completely consistent), determine the option sets of student A and student B respectively based on the option ranking, and then calculate the question type similarity corresponding to the multiple-choice question type between student A and student B based on the option sets of student A and student B.

[0198] That is, the set of options for student A ; Student B's option set

[0199] The question type similarity corresponding to the multiple-choice question type is calculated using the following formula:

[0200]

[0201] in, is the similarity coefficient corresponding to the multiple-choice question type; The question type similarity corresponding to the multiple-choice question type; A is the set of option answers for student A; B is the set of option answers for student B; |A∩B∣ indicates the number of options selected by both students, and |A∪B∣ indicates the number of options selected by at least one student.

[0202] Taking the text question type as an example, the present disclosure may regard the answers of student A and student B as a text set. For example, student A may give a text answer set A containing multiple keywords or phrases, while student B gives another text answer set B.

[0203] The question type similarity corresponding to the multiple-choice question type is calculated using the following formula:

[0204]

[0205] in, is the similarity coefficient corresponding to the text question type; is the question type similarity corresponding to the text question type; A is the text answer set of student A; B is the text answer set of student B; |A∩B| represents the number of common keywords or phrases in the two answer sets, and |A∪B| represents the number of all non-repeated keywords or phrases in the two answer sets.

[0206] It should be noted that the relative coefficient of the present disclosure has a value between 0 and 1: =0, it means that the two answer sets have no intersection, that is, student A and student B did not fill in any of the same options. =1, it means that the two answer sets are exactly the same, that is, student A and student B fill in exactly the same answers.

[0207] Step 703: Based on the number of students and the scores of each student under each question type, determine the score variance corresponding to each question type.

[0208] In some embodiments, the present disclosure can calculate the score variance corresponding to each question type based on the number of students and the scores of each student under each question type. The score variance reflects the degree of dispersion of the student's scores under this question type.

[0209] Among them, the score variance It can be calculated by the following formula:

[0210] is the student score of the i-th student in a certain question type; It is the mean (average) of students' scores in a certain answer type; is the number of students in the question type (the total number of students).

[0211] Step 704: Determine the minimum score variance among the score variances corresponding to each question type, and use each score variance minimum value as the test difficulty coefficient corresponding to each question type.

[0212] In some embodiments, the minimum score variance is used as the difficulty coefficient, which actually reflects the consistency of students' scores on this question type. The smaller the variance, the closer or more consistent the students' performance on this question type is, which means that the difficulty of this question type is moderate or the question design is relatively reasonable. Therefore, the present disclosure determines the minimum value minF in the score variance corresponding to each question type, and uses the minimum value as the examination difficulty coefficient of this question type. The smaller the difficulty coefficient, the easier the question type is.

[0213] Step 705: For each question type, based on the question type similarity and the test difficulty coefficient, determine the cheating suspicion score corresponding to each question type.

[0214] In some embodiments, the present disclosure may calculate the cheating suspicion score corresponding to each answering question type based on the question type similarity and the test difficulty coefficient for each answering question type.

[0215] Specifically, the cheating suspicion score can be obtained through the following formula:

[0216]

[0217] Among them, q is the comprehensive test difficulty coefficient, n is the number of students, and minF is the test difficulty coefficient corresponding to each question type. is the question type similarity corresponding to each answer question type between the first subject and the suspected cheating subject; The weight of the score for the preset cheating suspicion.

[0218] Step 706: Based on the preset score weights and the cheating suspicion score corresponding to each question type, the test paper result analysis feature value is obtained.

[0219] In some embodiments, after obtaining the cheating suspicion score corresponding to each answer question type, the present disclosure may weightedly fuse the cheating suspicion score corresponding to each answer question type through the following formula to obtain the test paper result analysis feature value.

[0220] Taking the example of the number of question types in this disclosure as 3, three cheating suspicion scores are obtained , , .

[0221]

[0222] in, , , The cheating suspicion score corresponding to each question type is respectively, , , The preset score weights.

[0223] In summary, the present disclosure comprehensively compares the answer data of the first subject and the suspected cheater, including answers and question types, calculates the similarity between the two in each question type, and determines the test difficulty coefficient in combination with the variance of the scores of the student group in each question type, and then generates a suspected cheating score for each question type. Finally, according to the preset score weight, the test paper result analysis feature value is summarized, and objective answer paper technical support is provided, which can accurately evaluate the possibility of cheating of the suspected cheater and improve the accuracy and efficiency of cheating detection.

[0224] Corresponding to the above-mentioned examination cheating analysis method, the present invention also provides an examination cheating analysis device. Since the device embodiment of the present invention corresponds to the above-mentioned method embodiment, details not disclosed in the device embodiment can be referred to the above-mentioned method embodiment, and will not be repeated in the present invention.

[0225] Figure 8 A schematic diagram of the structure of a test cheating analysis device provided by an embodiment of the present disclosure, such as Figure 8 As shown, the device comprises: The offset determination unit 810 is used to obtain multiple frames of images in the video stream and determine the test paper offset direction of the first object in each frame of the multiple frames of images; An exposure monitoring unit 820, for determining whether there is an exposure risk on the test paper based on the test paper coordinates and the first subject's hand coordinates in each frame of the image; The object monitoring unit 830 is used to determine the suspected cheating object associated with the first object based on the deviation direction of the test paper if there is a risk of exposure of the test paper; Comprehensive analysis unit 840, used to determine the angle feature value, eye contact and writing feature value and test result analysis feature value corresponding to the suspected cheater; The scoring unit 850 is used to determine a comprehensive score based on the angle feature value, the eye expression and writing feature value, and the test result analysis feature value, and determine the degree of cheating suspicion of the suspected cheater according to the comprehensive score.

[0226] In some embodiments of the present disclosure, the offset determination unit 810 is used to: perform color segmentation and edge detection on each frame image respectively, extract the test paper area and the table area in each frame image; determine the distance between the test paper vertex coordinates of the test paper area and the edge center coordinates of the table area; determine the test paper offset result based on the distance and a preset safety threshold; determine the test paper offset direction based on the test paper offset result.

[0227] In some embodiments of the present disclosure, the exposure monitoring unit 820 is used to: determine the overlapping vertex coordinates of the overlapping area between the test paper area and the hand area based on the test paper vertex coordinates of the test paper area in each frame image and the hand vertex coordinates of the hand area of ​​the first object; if the overlapping vertex coordinates meet the preset overlapping conditions, determine the overlapping area of ​​the overlapping area and the test paper area of ​​the test paper area based on the overlapping vertex coordinates and the test paper vertex coordinates; if the ratio between the overlapping area and the test paper area is less than the preset proportion threshold, determine that the test paper has a preliminary exposure risk; determine the overlapping ratio based on the hand vertex coordinates and the test paper vertex coordinates; if the overlapping ratio is greater than or equal to the preset overlapping ratio, determine that the test paper has an exposure risk.

[0228] In some embodiments of the present disclosure, the object monitoring unit 830 is used to: determine the target analysis population corresponding to the deviation direction of the test paper; and determine the suspected cheating objects in the target analysis population using the head posture of each target analysis object in the target analysis population.

[0229] In some embodiments of the present disclosure, the object monitoring unit 830 is used to: determine the associated population corresponding to the first object; determine the bird's-eye view angle of each object in the associated population looking at the first object based on the offset direction of the test paper and the position coordinates of each object in the associated population; and determine the target analysis population in the associated population based on the bird's-eye view angle.

[0230] In some embodiments of the present disclosure, the object monitoring unit 830 is used to: determine the head offset angle of each target analysis object based on the key point coordinates of each target analysis object; determine the relative angle between the first object and each target analysis object based on the test paper center point coordinates of each target analysis object and the position coordinates of the first object; determine the angle ratio between the head offset angle and the relative angle of each target analysis object, and determine the suspected cheating object in the target analysis population based on the angle ratio.

[0231] In some embodiments of the present disclosure, the comprehensive analysis unit 840 is used to: determine the horizontal deflection angle and the vertical pitch angle of the head of the first object, the seat angle between the first object and the suspected cheating object, and the seat pitch angle between the head of the first object and the desktop of the suspected cheating object; determine the horizontal angle score based on the seat angle between the first object and the suspected cheating object and the horizontal deflection angle of the head of the first object; determine the vertical angle score based on the vertical pitch angle of the head of the first object and the seat pitch angle between the head of the first object and the desktop of the suspected cheating object; determine the total angle score based on the horizontal angle score, the vertical angle score and the preset angle weight; determine the average angle score based on the total angle score and the number of frames corresponding to the video stream; determine the angle standard deviation based on the total angle score and the angle average score; determine the angle standard deviation based on the preset angle standard deviation threshold and the angle average score and the preset angle average score threshold, and obtain the angle feature value based on the comparison result.

[0232] In some embodiments of the present disclosure, the comprehensive analysis unit 840 is used to: determine the time that the suspected cheater stays in the writing area within a preset monitoring time period, where the writing area stay time includes the gaze stay time and the pen tip stay time; determine the ratio between the writing area stay time and the preset monitoring time period to obtain the writing area stay time ratio; compare the writing area stay time ratio and the preset stay time ratio threshold, and determine the eye writing feature value based on the comparison result.

[0233] In some embodiments of the present disclosure, the comprehensive analysis unit 840 is used to: obtain answer data of the first object and the suspected cheating object, the answer data including answer answers and answer question types; determine the question type similarity corresponding to each answer question type between the first object and the suspected cheating object based on the answer answers corresponding to each answer question type; determine the score variance corresponding to each answer question type based on the number of students and the student scores of each student under each answer question type; respectively determine the minimum score variance in the score variance corresponding to each answer question type, and use each minimum score variance as the exam difficulty coefficient corresponding to each answer question type; for each answer question type, determine the suspected cheating score corresponding to each answer question type based on the question type similarity and the exam difficulty coefficient; and obtain the test paper result analysis feature value based on the preset score weight and the suspected cheating score corresponding to each answer question type.

[0234] In some embodiments of the present disclosure, the evaluation unit 850 is used to: determine a comprehensive score based on angle feature values, eye contact and writing feature values, test paper result analysis feature values, and preset feature weights; if the comprehensive score is greater than or equal to a preset comprehensive score threshold, determine that the suspected cheating object has a high degree of suspected cheating; if the comprehensive score is less than the preset comprehensive score threshold, determine that the suspected cheating object has a low degree of suspected cheating.

[0235] It should be noted that the above explanation of the method embodiment is also applicable to the device of this embodiment, and the principle is the same, which is not limited in this embodiment.

[0236] Based on the above Figures 1 to 7 The method shown in the figure, accordingly, this embodiment also provides a computer program product, including a computer program, which implements the above-mentioned Figures 1 to 7 The method shown.

[0237] Based on the above Figures 1 to 7 The method shown in the embodiment also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the above-mentioned Figures 1 to 7 The method shown.

[0238] Based on this understanding, the technical solution of the present application can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, USB flash drive, mobile hard disk, etc.), including a number of instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute the methods of various implementation scenarios of the present application.

[0239] like Fig. 9 The figure shows a hardware structure diagram of an electronic device of the present invention, including: at least one processor 901; and, A memory 902 that is communicatively connected to at least one of the processors 901; wherein: The memory 902 stores instructions that can be executed by at least one of the processors. The instructions are executed by at least one of the processors so that the at least one of the processors can execute the test cheating analysis method as described above.

[0240] Fig. 9 A processor 901 is taken as an example.

[0241] The electronic device may further include: an input device 903 and a display device 904 .

[0242] The processor 901, the memory 902, the input device 903 and the display device 904 may be connected via a bus or other means, and the figure takes the connection via a bus as an example.

[0243] The memory 902 is a non-volatile computer-readable storage medium that can be used to store non-volatile software programs, non-volatile computer executable programs and modules, such as program instructions / modules corresponding to the review content generation method in the embodiment of the present application, for example, Figures 1 to 7 The processor 901 executes various functional applications and data processing by running the non-volatile software programs, instructions and modules stored in the memory 902, that is, the method for analyzing cheating in an examination in the above embodiment is implemented.

[0244] The memory 902 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function; the data storage area may store data created according to the use of the review content generation method, etc. In addition, the memory 902 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage devices. In some embodiments, the memory 902 may optionally include a memory remotely arranged relative to the processor 901, and these remote memories may be connected to the device for executing the review content generation method via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0245] The input device 903 can receive input user clicks and generate signal inputs related to user settings and function controls of the review content generation method. The display device 904 can include display devices such as a display screen.

[0246] The one or more modules are stored in the memory 902 and when executed by the one or more processors 901, the examination cheating analysis method in any of the above method embodiments is executed.

[0247] Optionally, the above-mentioned physical device may also include a user interface, a network interface, a camera, a radio frequency (RF) circuit, a sensor, an audio circuit, a WI-FI module, etc. The user interface may include a display, an input unit such as a keyboard, etc., and the optional user interface may also include a USB interface, a card reader interface, etc. The network interface may optionally include a standard wired interface, a wireless interface (such as a WI-FI interface), etc.

[0248] Those skilled in the art will appreciate that the above-mentioned physical device structure provided in this embodiment does not constitute a limitation on the physical device, and may include more or fewer components, or a combination of certain components, or different arrangements of components.

[0249] The storage medium may also include an operating system and a network communication module. The operating system is a program that manages the hardware and software resources of the above-mentioned physical device, and supports the operation of the information processing program and other software and / or programs. The network communication module is used to realize the communication between the components inside the storage medium, and the communication with other hardware and software in the information processing physical device.

[0250] Through the description of the above implementation methods, the technicians in this field can clearly understand that the present application can be implemented by means of software plus the necessary general hardware platform, or by hardware. By applying the scheme of this embodiment, compared with the current prior art, this embodiment obtains multiple frames of images in the video stream, and determines the test paper offset direction of the first object in each frame of the multiple frames; based on the test paper coordinates in each frame of the image and the hand coordinates of the first object, it is judged whether the test paper has an exposure risk; if the test paper has an exposure risk, the cheating suspect object associated with the first object is determined based on the test paper offset direction; the angle feature value, eye writing feature value and test paper result analysis feature value corresponding to the cheating suspect object are determined; based on the angle feature value, eye writing feature value and test paper result analysis feature value, the comprehensive score is determined, and the degree of cheating suspicion of the cheating suspect object is determined according to the comprehensive score. The use of intelligent video analysis technology, combined with the correlation analysis of the answer sheet results, greatly improves the efficiency and accuracy of cheating analysis, not only reduces the burden of manual review, improves the accuracy of cheating behavior identification, but also provides objective and sufficient cheating behavior evidence for educational institutions, which is convenient for subsequent processing and investigation.

[0251] It should be noted that, in this article, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0252] The above is only a specific implementation of the present application, so that those skilled in the art can understand or implement the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments described herein, but will conform to the widest scope consistent with the principles and novel features applied for herein.

Claims

1. A method for analyzing cheating in an examination, characterized in that: The method comprises: Acquire multiple frames of images in a video stream, and determine the test paper offset direction of the first object in each frame of the multiple frames of images; Based on the test paper coordinates and the hand coordinates of the first subject in each frame of the image, determining whether the test paper has an exposure risk; If the test paper has an exposure risk, determining a suspected cheating object associated with the first object based on the deviation direction of the test paper; Determine the angle feature value, eye contact and writing feature value, and test paper result analysis feature value corresponding to the suspected cheater; Based on the angle feature value, the eye contact and writing feature value and the test paper result analysis feature value, a comprehensive score is determined, and the degree of cheating suspicion of the cheating suspect is determined according to the comprehensive score.

2. The method according to claim 1, characterized in that The step of acquiring multiple frames of images in a video stream and determining the test paper offset direction of the first object in each frame of the multiple frames of images includes: Performing color segmentation and edge detection on each frame of the image respectively, and extracting the test paper area and the table area in each frame of the image; Determine the distance between the test paper vertex coordinates of the test paper area and the edge center coordinates of the table area; Based on the distance and a preset safety threshold, determining a test paper deviation result; Based on the test paper offset result, determine the test paper offset direction.

3. The method according to claim 1, characterized in that The step of judging whether the test paper has an exposure risk based on the test paper coordinates and the hand coordinates of the first object in each frame of the image includes: Based on the test paper vertex coordinates of the test paper area in each frame of the image and the hand vertex coordinates of the hand area of ​​the first object, determining the overlapping vertex coordinates of the overlapping area between the test paper area and the hand area; If the overlapping vertex coordinates meet the preset overlapping conditions, determining the overlapping area of ​​the overlapping region and the test paper area of ​​the test paper region based on the overlapping vertex coordinates and the test paper vertex coordinates; If the ratio between the overlapping area and the test paper area is less than a preset ratio threshold, it is determined that the test paper has a preliminary exposure risk; Determine an overlap ratio based on the hand vertex coordinates and the test paper vertex coordinates; If the overlap ratio is greater than or equal to a preset overlap ratio, it is determined that there is an exposure risk in the test paper.

4. The method according to claim 1, characterized in that: The step of determining the suspected cheating object associated with the first object based on the deviation direction of the test paper includes: Determine the target analysis population corresponding to the test paper deviation direction; The head posture of each target analysis object in the target analysis population is used to determine the suspected cheating object in the target analysis population.

5. The method according to claim 4, characterized in that The target analysis population corresponding to the test paper deviation direction is determined to include: Determine the associated population corresponding to the first object; Determine, according to the test paper offset direction and the position coordinates of each object in the associated group, a bird's-eye view angle of each object in the associated group looking at the first object; According to the overhead viewing angle, a target analysis population in the associated population is determined.

6. The method according to claim 4, characterized in that The step of using the head posture of each target analysis object in the target analysis population to determine the suspected cheating object in the target analysis population includes: Determining the head offset angle of each target analysis object according to the key point coordinates of each target analysis object; Determine the relative angle between the first object and each target analysis object according to the test paper center point coordinates of each target analysis object and the position coordinates of the first object; The angle ratio between the head deviation angle of each target analysis object and the relative angle is determined, and based on the angle ratio, the suspected cheating object in the target analysis population is determined.

7. The method according to claim 1, characterized in that Determining the angle feature value corresponding to the suspected cheating object includes: Determine a horizontal deflection angle and a vertical pitch angle of the head of the first subject, a seat angle between the first subject and the suspected cheater, and a seat pitch angle between the head of the first subject and a desk of the suspected cheater; Determining a horizontal angle score based on a seat angle between the first subject and the suspected cheater and a horizontal deflection angle of the first subject's head; Determining a vertical angle score based on the vertical pitch angle of the first subject's head and the seat pitch angle between the first subject's head and the tabletop of the suspected cheater; Determine a total angle score based on the horizontal angle score, the vertical angle score and a preset angle weight; Determine an average angle score based on the total angle score and the number of frames corresponding to the video stream; Determining the angle standard deviation based on the angle total score and the angle average score; The angle standard deviation is compared with a preset angle standard deviation threshold and the angle average score is compared with a preset angle average score threshold, and an angle feature value is obtained based on the comparison result.

8. The method according to claim 1, characterized in that Determining the eye contact writing feature value corresponding to the suspected cheater includes: Determine the time that the suspected cheater stays in the writing area within a preset monitoring time period, where the time that the suspected cheater stays in the writing area includes the time that the eyes stay and the time that the pen tip stays; Determine the ratio between the residence time in the writing area and the preset monitoring time period to obtain the residence time ratio in the writing area; The dwell time ratio of the writing area is compared with a preset dwell time ratio threshold, and the eye contact writing feature value is determined according to the comparison result.

9. The method according to claim 1, characterized in that: Determining the test paper result analysis feature value corresponding to the suspected cheater includes: Acquire answer data of the first subject and the suspected cheater, the answer data including answer and question type; Determining the question type similarity between the first object and the suspected cheating object for each question type based on the answer to each question type; Based on the number of students and the scores of each student under each question type, determine the score variance corresponding to each question type; Determine the minimum score variance among the score variances corresponding to each answering question type, and use each score variance minimum value as the test difficulty coefficient corresponding to each answering question type; For each question type, based on the question type similarity and the test difficulty coefficient, determine the cheating suspicion score corresponding to each question type; Based on the preset score weights and the cheating suspicion scores corresponding to each question type, the test paper result analysis feature values ​​are obtained.

10. The method according to claim 1, characterized in that The step of determining a comprehensive score based on the angle feature value, the eye contact and writing feature value, and the test paper result analysis feature value, and determining the degree of cheating suspicion of the cheating suspect according to the comprehensive score, includes: Determine a comprehensive score based on the angle feature value, the eye contact and writing feature value, the test paper result analysis feature value, and a preset feature weight; If the comprehensive score is greater than or equal to a preset comprehensive score threshold, the cheating suspicion level of the suspected cheating object is determined to be high cheating suspicion; If the comprehensive score is less than the preset comprehensive score threshold, it is determined that the cheating suspicion level of the cheating suspect is low cheating suspicion.

Images