Eight-section brocade intelligent scoring system based on machine learning and sum-product method

Abstract

The invention relates to a three-camera linkage and intelligent evaluation eight-brocade intelligent scoring system based on machine learning and a sum-product method. In an eight-section brocade scoring scene, a traditional subjective scoring mode has many defects. In order to solve the problems, an intelligent scoring system is constructed. The system comprises a video data acquisition module, a posture recognition module, an intelligent scoring module, a score integration module and the like. The video data acquisition module acquires multi-angle videos through three high-definition cameras; the posture recognition module accurately recognizes actions by using an advanced algorithm and a deep neural network technology; the intelligent scoring module designs a multi-dimensional index, and scores based on a sum-product algorithm; and the score integration module gathers scores from different perspectives to obtain a comprehensive score. According to the method, various measures are adopted to overcome the limitation of the existing AI technology in sports detection, including environment optimization and action detail identification, construction of a diversity and balance data set, research of cross-group adaptability, establishment of a man-machine cooperation trust relationship and the like. The system improves objectivity, accuracy and fairness of eight-segment brocade scoring, and is of great significance to eight-segment brocade teaching and competition. Meanwhile, a new technical means is provided for the fields of physical education and detection, the application range of the artificial intelligence technology in the field of physical education is expanded, and fusion innovation of the AI technology and the traditional field is promoted.

Description

Technical Field

[0001] The present invention relates to the application of artificial intelligence technology in the field of sports scoring, and in particular to the application of machine learning, posture recognition technology and sum-product method in the Ba Duan Jin intelligent scoring system, which belongs to the cross-technical field of artificial intelligence, computer vision and sports education. Background Art

[0002] 1. Limitations of traditional scoring methods In high school physical education courses, Ba Duan Jin is a compulsory subject and there are competitions every year. However, the traditional Ba Duan Jin scoring method relies heavily on the subjective judgment of the judges. During the scoring process, the judges' personal bias, experience differences and other factors make it difficult to ensure the fairness and accuracy of the scoring. This subjective scoring method cannot objectively reflect the true level of the contestants, and it is difficult for different judges to have completely consistent scoring standards, which to some extent affects the fairness of the Ba Duan Jin competition and the effectiveness of physical education teaching.

[0003] 2. Challenges of AI technology in sports testing and shortcomings of existing research 1. Sensitivity to environmental factors - When AI technology is currently applied to sports detection, environmental factors have a great impact on the recognition accuracy of the model. Taking the scoring system of Ba Duan Jin as an example, although the system performs well in an ideal environment, light changes and object occlusion are difficult to avoid in the actual sports competition environment. The competition environment is uncontrollable, and these variables will reduce the reliability of the AI ​​scoring system, causing the system's scoring results in different environments to fluctuate greatly.

[0004] 2. Movement details and complexity - The movements of Ba Duan Jin contain some subtle and complex posture changes, and it is difficult for AI systems to accurately identify these subtle movements. Existing AI systems may not be able to accurately capture these subtle movements, which will affect the comprehensiveness of the scoring and cause the scoring results to not accurately reflect the actual performance of the contestants in the Ba Duan Jin movements.

[0005] 3. Limitations of the dataset - The currently constructed dataset is insufficient in terms of the diversity and balance of movement samples. The lack of sufficient representative data limits the generalization ability of the model. When faced with unseen movements or styles, the model's scoring accuracy is greatly reduced, and it is impossible to accurately evaluate different types of Ba Duan Jin movements.

[0006] 4. Fairness and adaptability of the system - In sports testing, fairness needs to consider the adaptability of different groups of people (such as different ages, genders, and physical conditions). The current AI scoring system may not be fully verified in certain specific groups, resulting in limited application in various sports scenarios and unable to meet the needs of fair competition for all participants.

[0007] 5. Human-machine collaboration and trust - The consistency of AI scoring and human refereeing standards is an important issue. If there are differences in standards between the two, the acceptance of the scoring results by the contestants will be reduced. This problem will hinder the full promotion and acceptance of the AI ​​scoring system in sports testing and affect the in-depth application of AI technology in the field of sports. Summary of the invention

[0008] 1. System overall architecture The "three-camera linkage and intelligent evaluation" system of the present invention mainly includes the following modules: 1. Video data acquisition module - A three-camera linkage system is formed using three high-definition cameras to collect video data of Ba Duan Jin practitioners from different angles (such as the front, side, and upper side). These cameras can ensure that comprehensive Ba Duan Jin movement information is collected, and the collected video stream is transmitted to the PC through the network for subsequent processing. At the same time, privacy protection measures are strengthened during the collection process to ensure the information security of participants.

[0009] 2. Gesture Recognition Module - Apply advanced posture detection algorithms and integrate deep neural network technology in machine learning. First, use the public Ba Duan Jin dataset to train the model so that it can accurately identify Ba Duan Jin movements. In actual applications, this module analyzes the video data obtained from the video data acquisition module to achieve accurate positioning and tracking of Ba Duan Jin movements, extract action sequences for competition scoring, and effectively improve the accuracy of action recognition in the video.

[0010] 3. Intelligent scoring module - Design a multi-dimensional scoring index to comprehensively consider factors such as the standardization, fluency and completeness of the Ba Duan Jin movements. For a single video, conduct an in-depth study of the relationship between video frames and scores, especially the impact of repetitive video frames on the scores. On this basis, an algorithm based on the sum-product principle is used to score Ba Duan Jin practitioners according to the results of the posture recognition module. In the sum-product principle algorithm, different weight factors are set for different dimensions such as posture, rhythm and fluency. The scores of each dimension are calculated based on the movement recognition results and each weight factor, thereby reducing the reliance on subjective judgment and ensuring the comprehensiveness and fairness of the scoring.

[0011] 4. Rating Integration Module - The scoring results corresponding to cameras with different viewing angles are summarized through the sum-product method. During the summary process, the corresponding sum-product method coefficient is set according to the importance of each viewing angle, and the comprehensive score is finally output.

[0012] 2. Measures to overcome existing limitations 1. Optimization of environment and action detail recognition - To enhance the robustness of the model in complex environments, a variety of data enhancement techniques are used to simulate different lighting and occlusion conditions to train the model, so that the model can adapt to the light changes and possible object occlusion in various actual competition environments. At the same time, the posture detection algorithm is continuously optimized to improve the recognition ability of the subtle movements of Ba Duan Jin. By increasing the network depth and improving the feature extraction method, it is ensured that the subtle posture changes in Ba Duan Jin can be accurately captured.

[0013] 2. Construction of Diverse and Balanced Datasets - Collect more extensive and diverse Ba Duan Jin movement data, including movement data of practitioners of different ages, genders, and physical conditions, as well as Ba Duan Jin movement samples of different styles and difficulty levels. Classify and annotate the collected data, and construct the data set in a certain proportion to ensure the diversity and balance of the data set, thereby improving the generalization ability and fairness of the model.

[0014] 3. Cross-group adaptability research - Adjust and optimize the system according to the characteristics of different groups of people (such as different ages, genders, and physical conditions). By testing and verifying on sample data from different groups, adjust the scoring criteria and model parameters to ensure that the system performs consistently across different groups and that the scoring criteria are consistent and fair.

[0015] 4. Human-machine collaboration and trust building - To enhance the collaboration between the AI ​​system and human judges, the scoring logic is made transparent in the system design. Explain to users (including contestants, judges, etc.) how the scoring results are obtained, such as showing the scoring basis for each dimension and the calculation process of the sum-product method. In this way, users' trust in the AI ​​scoring results is enhanced, and contestants' acceptance of the scoring results is improved.

[0016] IV. Significance of the Invention 1. The significance of scoring Ba Duan Jin 1. Improve the objectivity and accuracy of scoring - The intelligent scoring system of the present invention is based on machine learning and the sum-product method, which avoids the interference of subjective factors of judges in traditional scoring methods. Through precise posture recognition technology and scientific scoring algorithms, it can conduct a comprehensive and detailed evaluation of Ba Duan Jin movements, thereby providing more objective and accurate scoring results, and more truly reflecting the Ba Duan Jin level of the contestants.

[0017] 2. Ensure fairness of scoring - The system collects data through multi-view cameras, analyzes the Ba Duan Jin movements from multiple angles, and adopts a scoring integration mechanism based on the sum-product method to comprehensively consider the movement performance from different angles. At the same time, in the scoring process, reasonable weight factors are set for different dimensions, taking into account all aspects of the movement, reducing the scoring deviation caused by a single factor, and ensuring the fairness of the scoring.

[0018] 2. Significance to the field of physical education and sports testing 1. Improve the level of modernization of physical education - In high school physical education, Ba Duan Jin is an important teaching content. The intelligent scoring system of the present invention provides a modern evaluation tool for Ba Duan Jin teaching. Teachers can use the system to more scientifically evaluate students' learning outcomes, timely identify students' problems in Ba Duan Jin practice and give targeted guidance, which is helpful to improve the quality and efficiency of Ba Duan Jin teaching and promote physical education to develop in an intelligent and scientific direction.

[0019] 2. Provide new technical means for sports testing - In the field of sports detection, especially in the scoring of traditional fitness events, this invention provides an innovative solution. It overcomes many limitations of existing AI technology in sports detection, provides a reference for the construction of intelligent scoring systems for other traditional sports, helps promote technological innovation in the entire sports detection field, and improves the fairness, accuracy and efficiency of sports competition scoring.

[0020] 3. Significance for the application of artificial intelligence technology 1. Expand the application scope of AI technology in sports - This invention demonstrates the successful application of artificial intelligence technology in the specific sports scenario of Ba Duan Jin scoring, and provides a model for the application of AI technology in other sports. By solving the problems of environmental factors, movement details, data set limitations, fairness and human-computer collaboration faced in sports detection, it provides technical ideas and practical experience for AI technology to further penetrate the sports field, which helps to expand the scope of application of AI technology in the sports field.

[0021] 2. Promote the integration and innovation of AI technology and traditional fields - Ba Duan Jin is a traditional fitness program. This invention integrates artificial intelligence technology with it, reflecting the innovative combination of traditional fields and modern technology. This integration not only injects new vitality into traditional sports, but also provides new application scenarios and demand drivers for the development of artificial intelligence technology, and promotes the integration and innovation of AI technology and traditional fields in more aspects. BRIEF DESCRIPTION OF THE DRAWINGS Figure 1 This is the physical construction diagram of the Ba Duan Jin classification system, which shows the connection relationship between the various components of the system and the general process of the system operation.

Claims

1. An intelligent scoring system for Ba Duan Jin based on machine learning and sum-product method, characterized in that: include: - Video data acquisition module, used to collect multi-angle video data of Ba Duan Jin practitioners through three cameras; - a motion recognition module, which uses deep learning technology to analyze the multi-angle video data to identify the movements of the Ba Duan Jin practitioner, wherein the movement analysis includes posture, rhythm and fluency; - Intelligent scoring module, which uses an algorithm based on the principle of sum-product method to score Ba Duan Jin practitioners according to the results of the action recognition module, so as to reduce the reliance on subjective judgment and ensure the comprehensiveness and fairness of the scoring; - The scoring integration module aggregates the scoring results corresponding to cameras with different viewing angles through the sum-product method to output a comprehensive score; - The system uses OpenCV and XEdu tools to integrate action recognition with dataset construction, model training and optimization, and scoring mechanism.

2. The Ba Duan Jin intelligent scoring system based on machine learning and sum-product method according to claim 1 is characterized in that: The three cameras in the video data acquisition module are respectively set at the front, side and upper angle to ensure that comprehensive Ba Duan Jin movement information is collected.

3. The Ba Duan Jin intelligent scoring system based on machine learning and sum-product method according to claim 1 is characterized in that: The deep learning technology in the action recognition module uses a convolutional neural network (CNN) algorithm to extract and recognize features of Ba Duan Jin actions.

4. The Ba Duan Jin intelligent scoring system based on machine learning and sum-product method according to claim 1 is characterized in that: In the sum-product principle algorithm of the intelligent scoring module, different weight factors are set for posture, rhythm and fluency respectively, and the score of each dimension is calculated based on the action recognition result and each weight factor.

5. The Ba Duan Jin intelligent scoring system based on machine learning and sum-product method according to claim 1 is characterized in that: When summarizing the scoring results of different perspectives, the scoring integration module sets corresponding sum-product method coefficients according to the importance of each perspective.

6. The Ba Duan Jin intelligent scoring system based on machine learning and sum-product method according to claim 1 is characterized in that: The model accuracy of the system reaches 97.16%, ensuring the accuracy of the scoring of Ba Duan Jin practitioners.

Images