A physical education examination system for assessing students' academic performance

Abstract

This invention discloses a physical education examination system for assessing student academic performance, belonging to the field of physical education examination systems. It includes: an image acquisition module equipped with multiple high-definition cameras arranged around a throwing circle to collect video data of the examinee during the throwing process; an identity verification module for performing two-stage identity verification; a foul detection module for analyzing video data to detect various fouls; a score processing module for handling score disputes; when there is an objection to the cancellation of a score due to a foul, the system retrieves the surveillance video, extracts a segment of the foul, enlarges it, and displays it on a large screen in the examination area to ensure fairness and transparency; and an output module for providing examinees with a valid score report after the examination, and separately listing areas for improvement. This invention, by using computer vision algorithms to replace or assist human judgment, eliminates the subjective bias of referees, making the identification of fouls more accurate and consistent, and greatly ensuring the fairness of the examination.

Description

Technical Field

[0001] This invention relates to the field of physical education examination systems, specifically a physical education examination system for assessing students' academic performance. Background Technology

[0002] Student academic proficiency examinations are a crucial link in measuring the quality of education and teaching and assessing students' comprehensive qualities. Physical education examinations, as an important component, have always received widespread attention regarding their fairness, accuracy, and efficiency. Shot put or medicine ball throwing is a common event in physical education entrance examinations for middle school, high school, and even university physical fitness tests. Its evaluation mainly relies on the measurement of throwing distance and strictly adheres to a series of rules to prevent cheating and ensure safety.

[0003] Currently, the traditional shot put distance measurement test process mainly relies on human referees and supervisors. Typically, a throwing circle or starting line is set up in the test area. One chief referee is responsible for issuing instructions, supervising the entire throwing process, and judging whether there are any fouls, while one or more assistant referees are responsible for measuring the landing distance, recording the score, and maintaining order on site.

[0004] However, traditional human referees and supervisors have the following problems: the judgment of fouls, such as "whether it started from a standstill" or "whether the foot stepped on the line," relies heavily on the referee's visual observation and instantaneous judgment; different referees may have blind spots, differences in reaction speed, and deviations in comprehension, leading to different judgments for the same action, affecting the absolute fairness and impartiality of the examination; especially in large-scale examinations, referees are prone to fatigue, leading to an increased rate of misjudgments and missed judgments. Therefore, those skilled in the art provide a physical education examination system for assessing students' academic performance to solve the problems mentioned in the background. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a physical education examination system for assessing student academic performance, comprising:

[0006] The image acquisition module is equipped with multiple high-definition cameras, which are arranged around the throwing circle to collect video data of the candidates during the throwing process. The video data during the throwing process focuses on the candidates' torso and feet.

[0007] The identity verification module is used to perform two-stage identity verification. The first stage uses dedicated sensors to perform facial recognition and ID card recognition. The second stage compares and analyzes the candidate's actions during the exam with the historical exam actions stored in the database. If the difference exceeds a predetermined threshold, an identity doubt warning is issued and deep identity verification is initiated, such as retrieving past exam records and asking the candidate about past scores.

[0008] The foul detection module is used to analyze video data to detect various foul behaviors, including whether the candidate starts throwing from a stationary position, whether any part of the body touches the ground outside the throwing circle or the starting line during the throwing process, whether the medicine ball is thrown correctly with both hands instead of using a swinging or throwing motion, whether the throw is completed within the specified time, and whether the candidate walks steadily from the back half of the throwing circle or behind the starting line after the throw. The determination of stability is obtained by comparing the walking posture when exiting with the initial walking posture when entering the examination room using pattern recognition technology.

[0009] The results processing module is used to handle results disputes. When there is an objection to the cancellation of results due to a foul, the monitoring video is retrieved, the foul action is extracted and enlarged and displayed on the large screen in the sports examination venue to ensure fairness and openness. In addition, if the system does not detect a foul but the distributed sound sensors capture the questioning sound, the video is deeply analyzed and manually identified to determine the validity of the results, and the analysis results and process are displayed on the large screen.

[0010] The output module is used to provide candidates with a valid score report after the exam, and to list the areas for improvement separately, such as technical defects such as incorrect launch angle leading to unsatisfactory distance or insufficient launch height leading to premature landing. These are based on video analysis and are designed to help improve future performance.

[0011] As a further aspect of the present invention: the high-definition camera of the image acquisition module has a resolution of at least 1080p and acquires video at a rate of more than 30 frames per second. The high-definition cameras are arranged in front of, to the sides and behind the throwing circle to form a multi-angle coverage network to ensure that the movement of the examinee's torso and feet is captured without blind spots. Each high-definition camera is equipped with autofocus and optical image stabilization to adapt to changes in outdoor light and motion blur. The video data is transmitted in real time to the foul detection module for synchronous analysis. The high-definition camera also integrates infrared night vision function to ensure operational reliability under low light conditions.

[0012] As a further aspect of the present invention: the second stage of the identity verification module adopts a deep learning-based action recognition algorithm. By extracting the habitual action features of the candidate during the throwing process, including the hand holding posture, pre-swing amplitude and throwing rhythm, an action feature vector is generated and its similarity is calculated with the feature vector in the historical database. If the similarity score is lower than a preset threshold, the system automatically triggers an identity doubt warning and starts a deep verification process. Deep verification includes retrieving the candidate's past exam videos for manual review or verifying identity through voice question and answer, ensuring the reliability of the anti-proxy test mechanism.

[0013] As a further aspect of the present invention: the foul detection module specifically includes:

[0014] The motion detection submodule uses continuous video frame difference analysis technology to detect the candidate's body movement before throwing, and confirms whether it starts from a static state by calculating the pixel change rate.

[0015] The line-stepping detection submodule uses a computer vision edge detection algorithm to identify the white line boundary of the throwing circle in real time and monitor the coordinates of the candidate's feet. If the foot coordinates exceed the boundary range, a foul is judged.

[0016] The motion standard detection submodule tracks and analyzes the throwing motion through key points of the skeleton to ensure that the medicine ball is thrown with both hands and without any non-legal movements.

[0017] The time detection submodule has a built-in timer that starts counting from when the candidate enters the throwing circle. If the candidate does not throw within the specified time (e.g., 30 seconds), an alarm will be triggered.

[0018] The post-throw detection submodule analyzes the candidate's path and posture when leaving the throwing circle, and uses gait analysis technology to compare the initial walking posture. If the posture deviation exceeds the tolerance, it is judged as unstable.

[0019] As a further aspect of the present invention: the results processing module includes a video storage submodule, a video retrieval submodule, and a display control submodule. The video storage submodule uses cloud storage technology to save all exam video clips and indexes timestamps and candidate information; the video retrieval submodule allows for quick retrieval of videos of controversial events through a graphical interface, supporting slow motion and zoom functions; the display control submodule outputs the processed video to a large screen in the sports field, displaying foul actions or analysis results in high-definition format, while integrating a manual recognition interface to allow referees to input the final judgment, and the system records all operation logs for auditing and transparency.

[0020] As a further aspect of the present invention: the output module generates a personalized report, including a summary of exam results and details of areas for improvement. The areas for improvement are quantitatively calculated based on video analysis algorithms, such as calculating the throwing angle and height of the shot put through trajectory tracking, and providing specific suggestions such as "increase the throwing angle by 5 degrees to optimize the distance" after comparing with the ideal value. The report is sent to the examinee's mobile device or printed out in electronic format (such as PDF), and at the same time provides a visual chart to show the historical progress trend to help examinees develop a training plan.

[0021] As a further aspect of the present invention: the similarity calculation for the second stage verification uses the following formula:

[0022] Where S is the similarity score, ranging from 0 to 1, with a higher value indicating greater similarity; The feature vector of the current test action includes the coordinates of key body points and the speed of movement; The feature vectors of historical exam actions were retrieved from the database. This represents the vector dot product operation; and Let represent the Euclidean norms of the two vectors, used for normalization.

[0023] As a further aspect of the invention, it also includes a distributed sound sensor module. The distributed sound sensor module is deployed in the spectator area and referee area of ​​the sports test venue. It uses microphone array technology to capture ambient sound and integrates sound recognition algorithms to filter noise. When a specific keyword (such as "foul" or "question") is detected, the system automatically triggers the score review process. The sound data and video data are time-stamped to ensure the consistency of the analysis. The distributed sound sensor module also has a waterproof and dustproof design to adapt to outdoor environments.

[0024] As a further aspect of the present invention, the system also includes a central control module. This module runs all analysis algorithms based on a high-performance processor (such as GPU acceleration) and connects to an external cloud database. The database stores candidates' historical exam records, video data, and identity information, supports big data analysis and machine learning model training, and the system communicates with each module through a wireless network to achieve real-time monitoring and remote management. It also has a user interface for administrators to configure exam parameters and view system status, ensuring the automation and intelligence of the entire exam process.

[0025] As a further aspect of the present invention, it also includes a potential assessment and recommendation module. This module is configured to: calculate a predicted potential improvement score based on the actual performance and list of areas for improvement provided by the output module, wherein the number and severity of the areas for improvement are positively correlated with the predicted potential improvement score; add the actual performance score to the predicted potential improvement score to obtain a potential score representing the candidate's best possible future performance level; retrieve the candidate's academic performance data from the academic affairs system; if the average academic performance score is lower than a preset passing score, and the potential score is higher than a potential threshold set for physical education students, generate an analysis report suggesting the candidate pursue a physical education development path, and send this suggestion to the candidate through the system interface; if the candidate confirms and agrees to the suggestion through the interface, the module matches information on physical education training teachers with corresponding teaching expertise based on the technical defects recorded in the student's past exams in the database, and pushes the teacher list and contact information to the candidate.

[0026] The beneficial effects of this invention are reflected in:

[0027] 1. Improve fairness and accuracy: By replacing or assisting human judgment with computer vision algorithms, the subjective bias of referees is eliminated, and the identification of fouls is more accurate and consistent, which greatly ensures the fairness of the examination.

[0028] 2. Enhanced anti-cheating capabilities: The two-stage identity verification mechanism, especially the verification at the behavioral biometric level, constitutes double insurance and can effectively prevent proxy test-taking.

[0029] 3. Enhance the transparency and credibility of examinations: The score dispute resolution mechanism provides intuitive video evidence and an open analysis process, making the judgments based on evidence, resolving disputes quickly, and enhancing the authority and credibility of the examination organization.

[0030] 4. Expanding Educational Evaluation Functions: The system is not only an assessment tool but also a teaching aid. The provided "List of Areas for Improvement" realizes the transformation from "summative assessment" to "formational assessment" and "diagnostic assessment," truly playing the role of "promoting learning through testing."

[0031] 5. Improved examination efficiency: Automated judging, measurement, and recording reduce reliance on human labor, speed up the examination process, and are particularly suitable for large-scale standardized examinations, reducing organizational costs. Attached Figure Description

[0032] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the accompanying drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn to scale.

[0033] Figure 1 This is a structural block diagram of a physical education examination system used to assess students' academic performance. Detailed Implementation

[0034] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0035] As mentioned in the background section of this application, research has revealed the following problems with existing human referees and supervisors: the judgment of fouls, such as "whether it started from a standstill" or "whether the foot stepped on the line," relies heavily on the referee's visual observation and instantaneous judgment; different referees may have blind spots, differences in reaction speed, and deviations in understanding, which may lead to different judgments for the same action, affecting the absolute fairness and impartiality of the examination; especially in large-scale examinations, referees are prone to fatigue, leading to an increased rate of misjudgments and missed judgments.

[0036] To address the aforementioned shortcomings, this application discloses a physical education examination system for assessing students' academic performance. By using computer vision algorithms to replace or assist human judgment, it eliminates the subjective bias of referees, making the identification of fouls more accurate and consistent, and greatly ensuring the fairness of the examination.

[0037] The following will describe in detail, with reference to the accompanying drawings, how the solution of this application solves the above-mentioned technical problems.

[0038] Please see Figure 1 In this embodiment of the invention, a physical education examination system for assessing student academic performance includes: an image acquisition module equipped with multiple high-definition cameras arranged around a throwing circle to collect video data of the examinee during the throwing process, with the video data focusing on the examinee's torso and feet; an identity verification module for performing two-stage identity verification, wherein the first stage uses dedicated sensors for facial recognition and ID card recognition, and the second stage compares and analyzes the examinee's actions during the examination with historical examination actions stored in a database. If the difference exceeds a predetermined threshold, an identity doubt warning is issued and deep identity verification is initiated, such as retrieving past examination records and inquiring about the examinee's past scores; and a foul detection module for analyzing video data to detect various foul behaviors, including whether the examinee starts throwing from a stationary state, whether any part of the body touches the upper edge of the throwing circle or the ground outside the throwing line during the throwing process, and whether the medicine ball is correctly thrown with both hands instead of being flicked or thrown down. The system assesses several aspects of a student's performance, including whether the throw was executed within the allotted time, and whether the student walked steadily from the back half of the throwing circle or behind the starting line. Stability is determined by comparing the student's walking posture upon exiting the throwing circle with their initial walking posture upon entering the testing area using pattern recognition technology. The score processing module handles disputes. When there is disagreement regarding a score cancellation due to a foul, the system retrieves surveillance video, extracts the foul's action, and displays it on a large screen in the testing area to ensure fairness and transparency. Furthermore, if the system does not detect a foul but the distributed sound sensors capture dissenting voices, the system performs in-depth video analysis and manual verification to determine the validity of the score, displaying the analysis results and process on the large screen. The output module provides students with a valid score report after the exam and lists areas for improvement, such as technical defects like incorrect throwing angle leading to unsatisfactory distance or insufficient throwing height causing premature landing. These findings, based on video analysis, aim to help improve future performance. This application realizes the full-process automation and intelligent management of physical education examinations. It integrates computer vision, sensor technology and data analysis to build a comprehensive solution that integrates identity verification to prevent cheating, real-time automatic monitoring of violations, fair handling of score disputes and post-examination learning analysis and feedback. It fundamentally solves the problems of low efficiency, strong subjectivity, easy to cause disputes and lack of data guidance value in the traditional manual refereeing method, and ensures the fairness, impartiality, openness and efficiency of the examination.

[0039] In this embodiment, the high-definition cameras of the image acquisition module have a resolution of at least 1080p and capture video at a rate of more than 30 frames per second. These high-definition cameras are positioned in front of, to the sides of, and behind the throwing circle, forming a multi-angle coverage network to ensure no blind spots in capturing the movements of the examinee's torso and feet. Each high-definition camera is equipped with autofocus and optical image stabilization to adapt to changes in outdoor light and motion blur. Video data is transmitted in real time to the foul detection module for synchronous analysis. The high-definition cameras also integrate infrared night vision to ensure operational reliability under low-light conditions. This image acquisition module provides the entire system with a high-quality, highly reliable, all-weather raw video data source. Its multi-angle, high frame rate, and high-resolution configuration ensures that subsequent analysis algorithms can accurately capture the examinee's most subtle movement details (such as whether the foot steps on the line or the posture at the moment the ball is released). The autofocus, image stabilization, and night vision functions ensure the stability and consistency of data acquisition under different environmental conditions, laying a solid data foundation for the implementation of all subsequent analysis functions.

[0040] In this embodiment, the second-stage verification of the identity verification module employs a deep learning-based action recognition algorithm. By extracting habitual action features from the candidate's throwing process, including hand-held ball posture, pre-swing amplitude, and throwing rhythm, an action feature vector is generated. This vector is then compared with feature vectors in the historical database. If the similarity score is lower than a preset threshold (e.g., 0.7), the system automatically triggers an identity doubt warning and initiates a deep verification process. Deep verification includes retrieving the candidate's past exam videos for manual review or verifying identity through voice question-and-answer sessions, ensuring the reliability of the anti-cheating mechanism. This second-stage identity verification module provides a dynamic, biometric-based identity verification method as an important supplement to static facial recognition. By analyzing each individual's unique habitual action patterns, the system can effectively identify those using counterfeit documents or those who have passed facial verification as substitute test-takers, greatly enhancing the anti-cheating capabilities of the examination system and addressing potential vulnerabilities arising from relying solely on initial identity verification.

[0041] In this embodiment, the foul detection module specifically includes: a start-of-motion detection submodule, which uses continuous video frame difference analysis technology to detect the candidate's body movement before throwing, and confirms whether it starts from a static state by calculating the pixel change rate; a line-stepping detection submodule, which uses a computer vision edge detection algorithm to identify the white line boundary of the throwing circle in real time and monitors the candidate's foot position coordinates. If the foot coordinates exceed the boundary range, a foul is judged; a motion standard detection submodule, which analyzes the throwing action through skeletal key point tracking to ensure that the medicine ball is thrown with both hands and there are no non-legal actions such as one-handed flicking or throwing downwards; a time detection submodule, which has a built-in timer that starts timing from when the candidate enters the throwing circle. If the ball is not released within the specified time (e.g., 30 seconds), an alarm is triggered; and a release detection submodule, which analyzes the candidate's path and posture when leaving the throwing circle, and uses gait analysis technology to compare the initial walking posture. If the posture deviation exceeds the tolerance, instability is judged. This foul detection module replaces human referees and realizes automated, refined, and real-time identification and judgment of various complex foul behaviors in the medicine ball test. By breaking down the overall foul rules into multiple computable sub-tasks and employing the most suitable computer vision algorithm for each sub-task, it achieves high-precision foul detection, significantly reduces misjudgments and missed judgments, ensures the objectivity and consistency of rulings, and greatly reduces the workload of referees.

[0042] In this embodiment, the score processing module includes a video storage submodule, a video retrieval submodule, and a display control submodule. The video storage submodule uses cloud storage technology to save all exam video clips and indexes timestamps and candidate information. The video retrieval submodule allows for quick retrieval of videos of controversial events through a graphical interface, supporting slow-motion and zoom-in functions. The display control submodule outputs the processed video to a large screen in the sports field, displaying foul actions or analysis results in high-definition format. It also integrates a human recognition interface, allowing referees to input their final judgments. The system records all operation logs for auditing and transparency. This score processing module constructs a transparent and traceable dispute resolution and arbitration mechanism. When score disputes arise, this module can quickly locate and present key evidence videos, making the basis for the judgment visible by publicly displaying foul details or the system analysis process, thereby effectively persuading candidates, resolving disputes, and maintaining the credibility of the exam. Simultaneously, its human-computer interaction design ensures the organic combination of automated system judgment and the referee's final decision-making power.

[0043] In this embodiment, the output module generates a personalized report, including a summary of exam results and details of areas for improvement. These areas for improvement are quantified using video analytics algorithms. For example, the throw angle and height of the medicine ball are calculated through trajectory tracking, and specific suggestions are given after comparison with ideal values, such as "increase the throw angle by 5 degrees to optimize the distance." The report is sent to the examinee's mobile device in electronic format (e.g., PDF) or printed out. It also provides visual charts to show historical progress trends, helping examinees develop training plans. This output module extends the exam system from a simple "evaluation tool" to a "teaching aid," providing each examinee with a quantitative and actionable personalized sports performance analysis report through data mining technology. It not only tells examinees "what their score is," but more importantly, answers "how they can do better," thus serving the fundamental purpose of improving students' physical education academic level and achieving a closed loop of testing and guidance.

[0044] In this embodiment, the throw angle is calculated using the following formula: Where θ is the launch angle of the solid sphere, in degrees; v x The horizontal velocity of the solid sphere is calculated from the rate of change of position in the video sequence; v y The vertical velocity of the medicine ball is estimated from video analysis; arctan is the arctangent function used to derive the angle from the velocity components. This formula, based on fundamental kinematic principles, converts the two-dimensional velocity data obtained from video analysis into the crucial parameter of the launch angle, which is vital for motion performance. This calculation provides precise data support for the "Improvements" section in the output module, enabling the system to specifically indicate the extent of "angle deviation," thus making the improvement suggestions (such as "increase the launch angle by 5 degrees") scientific and targeted, rather than vague descriptions.

[0045] In this embodiment, the similarity calculation for the second-stage verification uses the following formula: Where S is the similarity score, ranging from 0 to 1, with a higher value indicating greater similarity; The feature vector of the current test action includes the coordinates of key body points and the speed of movement; The feature vectors of historical exam actions were retrieved from the database. This represents the vector dot product operation; and Let and represent the Euclidean norms of the two vectors, used for normalization. This formula provides an objective and quantitative mathematical metric for comparing action similarity in identity verification. By calculating cosine similarity, this formula effectively ignores the interference of irrelevant factors such as action amplitude and absolute position, focusing on capturing the essential similarities and differences in action patterns in direction. This allows for accurate determination of whether the current action and historical records originate from the same examinee, providing a reliable numerical basis for identity-related warnings.

[0046] This embodiment also includes a distributed sound sensor module, which is deployed in the spectator and referee areas of the sports examination venue. It uses microphone array technology to capture ambient sound and integrates a sound recognition algorithm to filter noise. When specific keywords (such as "foul" or "question") are detected, the system automatically triggers a score review process. Sound data and video data are timestamped to ensure consistency in analysis. The distributed sound sensor module is also waterproof and dustproof, adaptable to outdoor environments. The sound sensor acts as a proactive, environmentally-aware "early warning trigger," expanding the system's perception dimension beyond visual analysis. By capturing the questioning sounds of on-site personnel, the system can discover potential points of contention that its own algorithm might have missed, thereby initiating the review mechanism. This reflects the redundancy in the system design and the pursuit of the ultimate fairness, further enhancing the reliability and public trust of the entire examination system.

[0047] In this embodiment, the system also includes a central control module. This module runs all analysis algorithms based on a high-performance processor (such as GPU acceleration) and connects to an external cloud database. The database stores candidates' historical exam records, video data, and identity information, supporting big data analysis and machine learning model training. The system communicates with each module via a wireless network to achieve real-time monitoring and remote management. It also has a user interface for administrators to configure exam parameters and view system status, ensuring the automation and intelligence of the entire examination process. This central control module acts as the "brain" and "central nervous system" of the entire system, responsible for coordinating, scheduling, and calculating the work of all modules. Its high-performance hardware ensures the real-time operation of complex AI algorithms; the cloud database enables the storage and mining of massive amounts of data; and wireless communication and the user interface provide the capabilities for system integration, remote operation, and human-computer interaction. Ultimately, it integrates the various disparate modules into a stable, efficient, and scalable intelligent whole.

[0048] In this embodiment, a potential assessment and recommendation module is also included. This module is configured to: calculate a predicted potential improvement score based on the actual performance and list of areas for improvement provided by the output module, wherein the number and severity of areas for improvement are positively correlated with the predicted potential improvement score; add the actual performance score to the predicted potential improvement score to obtain a potential score representing the candidate's best possible future performance level; retrieve the candidate's academic performance data from the academic affairs system; if the average academic performance score is lower than a preset passing score, and the potential score is higher than the potential threshold set for physical education students, generate an analysis report suggesting the candidate pursue a physical education development path and send this suggestion to the candidate through the system interface; if the candidate confirms and agrees to the suggestion through the interface, the module matches information on physical education training teachers with corresponding teaching expertise based on the technical deficiencies recorded in the student's past exams in the database, and pushes the teacher list and contact information to the candidate. This setup elevates a simple examination and testing system to the level of an academic and career planning support system, utilizing a large amount of data generated in the examination system (action data, performance data, historical deficiency data) to provide students with forward-looking and valuable decision support.

[0049] Compared to traditional human referees and supervisors, this invention offers the following advantages: 1. Enhanced fairness and accuracy: By replacing or assisting human judgment with computer vision algorithms, subjective bias by referees is eliminated, resulting in more accurate and consistent identification of violations and greatly ensuring the fairness of the examination. 2. Strengthened anti-cheating capabilities: The two-stage identity verification mechanism, especially at the behavioral biometric level, provides double insurance, effectively preventing proxy test-taking. 3. Increased transparency and credibility: The grade dispute resolution mechanism provides intuitive video evidence and a transparent analysis process, ensuring that penalties are based on evidence and disputes are quickly resolved, enhancing the authority and credibility of the examination organization. 4. Expanded educational evaluation functions: The system is not only an assessment tool but also a teaching aid. The provided "list of areas for improvement" shifts from "summative evaluation" to "formational evaluation" and "diagnostic evaluation," truly playing the role of "learning through testing." 5. Improved examination efficiency: Automated judgment, measurement, and recording reduce reliance on human labor, accelerate the examination process, and are particularly suitable for large-scale standardized examinations, reducing organizational costs.

[0050] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be covered within the scope of the claims and specification of the present invention.

Claims

1. A physical education examination system for assessing students' academic performance, characterized in that, include: The image acquisition module is equipped with multiple high-definition cameras, which are arranged around the throwing circle to collect video data of the candidates during the throwing process; The identity verification module is used to perform two-stage identity verification. The first stage uses dedicated sensors to perform facial recognition and ID card recognition. The second stage compares and analyzes the candidate's actions during the exam with the historical exam actions stored in the database. If the difference exceeds a predetermined threshold, an identity doubt warning is issued and deep identity verification is initiated. The foul detection module is used to analyze video data to detect various foul behaviors, including whether the candidate starts throwing from a stationary position, whether any part of the body touches the ground outside the throwing circle or the starting line during the throwing process, whether the medicine ball is thrown correctly with both hands instead of using a swinging or throwing motion, whether the throw is completed within the specified time, and whether the candidate walks steadily from the back half of the throwing circle or behind the starting line after the throw. The determination of stability is obtained by comparing the walking posture when exiting with the initial walking posture when entering the examination room using pattern recognition technology. The results processing module is used to handle results disputes. When there is an objection to the cancellation of results due to a foul, the monitoring video is retrieved, the foul action is extracted and enlarged and displayed on the large screen in the sports examination venue to ensure fairness and openness. In addition, if the system does not detect a foul but the distributed sound sensors capture the questioning sound, the video is deeply analyzed and manually identified to determine the validity of the results, and the analysis results and process are displayed on the large screen. The output module is used to provide candidates with a valid score report after the exam, and to separately list areas for improvement.

2. The physical education examination system for assessing student academic performance according to claim 1, characterized in that, The image acquisition module's high-definition cameras have a resolution of at least 1080p and capture video at a rate of more than 30 frames per second. The high-definition cameras are arranged in front of, to the sides of, and behind the throwing circle to form a multi-angle coverage network, ensuring that the movement of the examinee's torso and feet is captured without blind spots. Each high-definition camera is equipped with autofocus and optical image stabilization to adapt to changes in outdoor light and motion blur. The video data is transmitted in real time to the foul detection module for synchronous analysis. The high-definition cameras also integrate infrared night vision to ensure operational reliability under low-light conditions.

3. A physical education examination system for assessing student academic performance according to claim 2, characterized in that, The second stage of the identity verification module employs a deep learning-based action recognition algorithm. By extracting habitual action features from the candidate's throwing process, including the hand-holding posture, pre-swing amplitude, and throwing rhythm, an action feature vector is generated. This vector is then compared with feature vectors in the historical database to calculate similarity. If the similarity score is lower than a preset threshold, the system automatically triggers an identity doubt warning and initiates a deep verification process. Deep verification includes retrieving the candidate's past exam videos for manual review or verifying identity through voice Q&A, ensuring the reliability of the anti-proxy test mechanism.

4. A physical education examination system for assessing student academic performance according to claim 3, characterized in that, The foul detection module specifically includes: The motion detection submodule uses continuous video frame difference analysis technology to detect the candidate's body movement before throwing, and confirms whether it starts from a static state by calculating the pixel change rate. The line-stepping detection submodule uses a computer vision edge detection algorithm to identify the white line boundary of the throwing circle in real time and monitor the coordinates of the candidate's feet. If the foot coordinates exceed the boundary range, a foul is judged. The motion standard detection submodule tracks and analyzes the throwing motion through key points of the skeleton to ensure that the medicine ball is thrown with both hands and without any non-legal movements. The time detection submodule has a built-in timer that starts counting from when the candidate enters the throwing circle. If the candidate does not throw within the specified time, an alarm will be triggered. The post-throw detection submodule analyzes the candidate's path and posture when leaving the throwing circle, and uses gait analysis technology to compare the initial walking posture. If the posture deviation exceeds the tolerance, it is judged as unstable.

5. A physical education examination system for assessing student academic performance according to claim 4, characterized in that, The score processing module includes a video storage submodule, a video retrieval submodule, and a display control submodule. The video storage submodule uses cloud storage technology to save all exam video clips and indexes timestamps and candidate information. The video retrieval submodule allows for quick retrieval of videos of controversial events through a graphical interface, supporting slow-motion and zoom-in functions; The display control submodule outputs the processed video to the stadium's large screen, displaying the fouls or analysis results in high definition. It also integrates a human recognition interface, allowing referees to input their final judgment.

6. A physical education examination system for assessing student academic performance according to claim 5, characterized in that, The output module generates a personalized report, including a summary of exam results and details of areas for improvement. The areas for improvement are quantified based on video analytics algorithms. The report is sent electronically to the candidate's mobile device or printed out, and also provides visual charts to show historical progress trends to help candidates develop training plans.

7. A physical education examination system for assessing student academic performance according to claim 6, characterized in that, The similarity calculation for the second-stage verification uses the following formula: Where S is the similarity score, ranging from 0 to 1, with a higher value indicating greater similarity; The feature vector of the current test action includes the coordinates of key body points and the speed of movement; The feature vectors of historical exam actions were retrieved from the database. This represents the vector dot product operation; and Let represent the Euclidean norms of the two vectors, used for normalization.

8. A physical education examination system for assessing student academic performance according to claim 7, characterized in that, It also includes a distributed sound sensor module, which is deployed in the spectator and referee areas of the sports test site. It uses microphone array technology to capture ambient sound and integrates sound recognition algorithms to filter noise. When a specific keyword is detected, the system automatically triggers the score review process. The sound data and video data are time-stamped to ensure the consistency of the analysis. The distributed sound sensor module also has a waterproof and dustproof design to adapt to outdoor environments.

9. A physical education examination system for assessing student academic performance according to claim 8, characterized in that, The system also includes a central control module, which runs all analysis algorithms based on a high-performance processor and connects to an external cloud database. The database stores candidates' historical exam records, video data, and identity information, and supports big data analysis and machine learning model training. The system communicates with each module via a wireless network to achieve real-time monitoring and remote management. It also has a user interface for administrators to configure exam parameters and view system status, ensuring the automation and intelligence of the entire exam process.

10. A physical education examination system for assessing student academic performance according to claim 9, characterized in that, It also includes a potential assessment and recommendation module; the potential assessment and recommendation module is configured to: calculate a predicted performance improvement based on the actual performance and list of areas for improvement provided by the output module, wherein the number and severity of the areas for improvement are positively correlated with the predicted performance improvement value; The actual score is added to the predicted improvement score to obtain the potential score, which represents the candidate's best possible future performance. The candidate's academic performance data in the academic affairs system is retrieved. If the average academic score is lower than the preset passing score, and the potential score is higher than the potential threshold set for sports students, an analysis report suggesting that the candidate pursue a sports development path is generated and sent to the candidate through the system interface. If the candidate confirms and agrees to the suggestion through the interface, the module matches information on sports training teachers with corresponding teaching expertise based on the technical defects recorded in the student's past exams in the database, and pushes the list of teachers and their contact information to the candidate.

Images