Systematic training of table tennis players' physical performance based on artificial intelligence technology and data fusion of sensing devices

Abstract

This research emphasises the value of physical training for table tennis players, particularly as ball speed and spin rate decline and emphasises how important intensity quality is to the game. Chinese table tennis players' dual identities place greater demands on the general growth of their learning and training as a crucial component of talent development preparation. Athletes' general quality, competitive level, and ability to avoid sports injuries are all improved by scientific and focused physical training. In order to achieve the functions of intelligent camera, multi-angle broadcasting, and 3D scene reproduction, this study combines the physical training model of artificial intelligence. This gives the audience a more engaging and in-depth viewing experience. More feature extraction of the match footage is made possible by deep learning and convolutional neural networks when combined with large-scale video data, greatly enhancing the match information for viewers. The experimental findings demonstrate that the accuracy of table tennis human technical movement recognition reaches 98.88 % based on the enhanced AM-Softmax classification algorithm.