Design of the sport training assistant system based on local spatio-temporal features

In order to improve and strengthen the performance of the sports training assistant system, this paper designs a sports training assistant system based on the local spatio-temporal features. First, the coordinate information of human joints is obtained by using local space-time features; Then, the motion behavior is recognized by using the dynamic time warping algorithm based on the coordinate information; Finally, the matching degree between the model behavior and real-time behavior is evaluated and compared. The experimental results show that the system based on local spatiotemporal features has better recognition rate and evaluation rate, which can provide timely and effective feedback for sports evaluation.