The application of deep learning-based technique detection model in table tennis teaching and learning

Abstract

With the development of computer technology, the teaching methods of table tennis have ushered in a new technological revolution. To solve the problem of traditional teaching methods overly focusing on athlete limbs and athlete force movements, this study uses an improved deep learning algorithm technology detection model to analyze the trajectory of table tennis and provide targeted tactical training for athletes. The results showed that the success rate and accuracy score of the model were 95 % and 96 %, respectively, with a calculation time of only 21.75 ms, indicating high analytical accuracy and computational efficiency. Meanwhile, the winning rate of the training strategy under this method can reach over 65 %, effectively improving the winning rate of athletes. This proves that the proposed technology detection model has good algorithm performance and data analysis ability, and can provide data support for table tennis training and teaching work.

Keywords

• Deep learning
• Table tennis
• Trajectory tracking
• Hof circle detection
• Transfer learning