Design of Motor Skill Recognition and Hierarchical Evaluation System for Table Tennis Players

With the rapid development of electronic science and technology, the research on wearable devices is constantly updated. Still, the current wearable devices are not yet comprehensive in recognizing and analyzing the movement patterns of special sports. Based on this, this article improves wearable devices for table tennis, a single sport, and realizes the recognition and evaluation of table tennis players' motor skills based on the pattern recognition method and hierarchical evaluation system. First, a table tennis player's motion information acquisition device was designed and the collected motion data were processed. Second, a sliding window was designed to segment the collected motion data and establish a feature database of six table tennis benchmark movements. Third, the motion features were constructed based on feature engineering, and after dimensionality reduction, the different models were used for action skill recognition. Finally, a hierarchical evaluation system of movement skills was established and the loss functions of different evaluation indexes were given. The experimental results show that in table tennis players' action skill recognition, the neural network model based on motor feature extraction proposed in this article improves the recognition accuracy by 6.5% than the traditional convolutional neural network, and in table tennis players' action skill evaluation, the analysis of variance (ANOVA) shows that the hierarchical evaluation system proposed in this article has a significance level of p < 0.05 for the athletes' professionalism level. The conclusion of the study indicates that the proposed paper table tennis players' motor skills recognition and hierarchical evaluation assessment system can realize the accurate recognition of athletes' motor skills and the precise determination of their professional level.