Shi, Shaodong
(School of Computer and Information, Hefei University of Technology, Hefei, China)
,
Zhou, Qing F.
(School of Electrical Engineering & Intelligentization, Dongguan University of Technology, Dongguan, China)
,
Peng, M.
(School of Computer and Information, Hefei University of Technology, Hefei, China)
,
Cheng, Xusheng
(School of Computer and Information, Hefei University of Technology, Hefei, China)
Statistics on motions (e.g. turning around) in basketball games are important reference information for training after the games. In this paper, we propose a system that uses smart insole to recognize basketball motions. Relying on the built-in motion sensor in the insole, three-axis acceleration da...
Statistics on motions (e.g. turning around) in basketball games are important reference information for training after the games. In this paper, we propose a system that uses smart insole to recognize basketball motions. Relying on the built-in motion sensor in the insole, three-axis acceleration data and three-axis angular velocity data about three different basketball motions can be collected. These three basketball motions are dribble, jump, and turning around. The K-means clustering algorithm is utilized to complete the recognition task, and a new method is used to initialize the center points, which solves the problem of large difference between the results of each experiment. In the case of training data and test data from different players, the average accuracy can reach 98.9%.
Statistics on motions (e.g. turning around) in basketball games are important reference information for training after the games. In this paper, we propose a system that uses smart insole to recognize basketball motions. Relying on the built-in motion sensor in the insole, three-axis acceleration data and three-axis angular velocity data about three different basketball motions can be collected. These three basketball motions are dribble, jump, and turning around. The K-means clustering algorithm is utilized to complete the recognition task, and a new method is used to initialize the center points, which solves the problem of large difference between the results of each experiment. In the case of training data and test data from different players, the average accuracy can reach 98.9%.
※ AI-Helper는 부적절한 답변을 할 수 있습니다.