Personalized Intervention Research on University Table Tennis Training Based on Artificial Intelligence and Learning Analytics Technology

DOI

10.2991/978-94-6239-691-3_64How to use a DOI?

Keywords

Artificial Intelligence; Learning Analytics; Table Tennis Training; Pose Estimation; Spatial Temporal Graph Convolutional Networks; Personalized Intervention

Abstract

To address the problems of lack of personalized guidance, subjective action evaluation, and delayed feedback in traditional table tennis teaching in universities, this paper proposes and implements a personalized training intervention system integrating computer vision, deep learning, and learning analytics technology. The system employs the OpenPose pose estimation algorithm to extract coordinates of 14 upper limb keypoints from students in real time. Targeting the characteristics of fast table tennis movements and frequent occlusions, multi-scale feature fusion and temporal consistency constraints are introduced for optimization, improving keypoint detection accuracy from 86.3% to 91.7% while maintaining a detection frame rate above 25fps. On this basis, an action recognition model based on Spatial Temporal Graph Convolutional Networks (ST-GCN) is constructed. Through three optimizations—introduction of attention mechanisms, multi-scale temporal convolution kernels, and label smoothing regularization—the model achieves automatic classification of four fundamental movements (forehand drive, backhand push, forehand chop, and backhand chop) with a recognition accuracy of 92.7%, representing an 8.4 percentage point improvement over 3D-CNN. Furthermore, a continuous scoring model integrating three dimensions—keypoint position accuracy, joint angle matching degree, and motion trajectory smoothness—is designed, achieving a correlation coefficient of 0.81 between model scores and expert ratings with a root mean square error of 4.18. Finally, a learning analytics module based on bidirectional Long Short-Term Memory (Bi-LSTM) networks is introduced, combined with attention mechanisms to dynamically track student training trajectories, categorizing students into four learning stages and matching personalized intervention strategies. A quasi-experimental study was conducted in a university table tennis course, with 86 students randomly divided into experimental and control groups. After an 8-week teaching experiment, results show that students in the experimental group improved their forehand drive scores by 16.3 points and backhand push scores by 15.2 points, significantly higher than the 7.8-point and 7.3-point improvements in the control group (p < 0.001), while their skill acquisition cycle was shortened by 23.4%. The proposed method effectively realizes intelligent and personalized table tennis training, providing a reference technical pathway for the reform of university physical education.

Copyright

© 2026 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Yan Zheng
PY  - 2026
DA  - 2026/05/31
TI  - Personalized Intervention Research on University Table Tennis Training Based on Artificial Intelligence and Learning Analytics Technology
BT  - Proceedings of the 2026 5th International Conference on Educational Innovation and Multimedia Technology (EIMT 2026)
PB  - Atlantis Press
SP  - 635
EP  - 646
SN  - 2667-128X
UR  - https://doi.org/10.2991/978-94-6239-691-3_64
DO  - 10.2991/978-94-6239-691-3_64
ID  - Zheng2026
ER  -