Affiliations: [a]
Department of Industrial Design, School of Electromechanic Engineering, Qingdao University, Qingdao, China
| [b]
School of Sciences, Qingdao University of Technology, Qingdao, China
Correspondence:
[*]
Corresponding author. Fei Kong, Department of Industrial Design, School of Electromechanic Engineering, Qingdao University, Qingdao, China. E-mail: [email protected].
Abstract: At present, there is less software related to sport technical behavior recognition, and there are few studies on the classification and identification of detailed actions. By introducing computer technology to analyze the efficiency and regularity of sports, not only the characteristics of athletes can be excavated, but also the visibility and dynamic tracking of sport training can be provided. The process of sports education is a fast and complex systematic process. Through the interactive system of physical education, we can use different methods to collect sports data and make a comparative analysis of athletes’ movements. Through the data mining of the relationship between athletes’ physiological indexes and sports load, the unreasonable link in sports training can be avoided. Also, in sports training, we can use computer vision and modern biomechanics to construct a virtual sports education situation. With the classification accuracy as the fitness function, this paper collects the data through the network database, and returns the corresponding sport training parameters on this basis. The results showed that the accuracy of the model was nearly 98%, which met the actual demand. Therefore, the development of sports education assistant system can provide strong support for the process control of sports training and education.
Keywords: Support vector, improved model, computer interactive system, recognition algorithm
