SPACE-TIME GRAPH-BASED CONVOLUTIONAL NEURAL NETWORKS OF STUDY ON MOVEMENT RECOGNITION OF FOOTBALL PLAYERS

Authors

  • Siming Tian College of Humanities, Chongqing Metropolitan College of Science and Technology, Chongqing 402167, China
  • Bingcheng Yin PhD Candidate at South Seoul University, South Korea, Xingtai 053001, China
  • Lei Wang Department of Physical Education, Tangshan Normal University, Tangshan 063000, China
  • Li Sun Hebei Vocational College for Correctional Police

DOI:

https://doi.org/10.23055/ijietap.2023.30.2.8581

Abstract

Behaviour recognition technology is an interdisciplinary technology, integrating many research achievements in computer vision, deep learning, pattern recognition and other fields. The key information of bone data on human behavior can not only accurately describe the motion posture of the human body in three-dimensional space, but also its rigid connection structure is robust to various external interference factors. However, the behavioral recognition algorithm is influenced by different factors such as background, light and environment, which is easy to lead to unstable recognition accuracy and limited application scenarios. To address this problem, in this paper, we propose a noise filtering algorithm based on data correlation and skeleton energy model filtering, construct a set of football player data sets, using the ST-GCN algorithm to train the skeleton characteristics of football players, and construct a behavior recognition system applied to football players. Finally, by comparing the accuracy of Deep LSTM, 2s-AGCN and the algorithm in this paper, the accuracy of TOP1 and TOP5 is 39.97% and 66.34%, respectively, which are significantly higher than the other two algorithms. It can realize the statistics of athletes and analyze the technical and tactical movements of players on the football field.

Downloads

Published

2023-04-18

How to Cite

Siming Tian, Bingcheng Yin, Lei Wang, & Sun, L. (2023). SPACE-TIME GRAPH-BASED CONVOLUTIONAL NEURAL NETWORKS OF STUDY ON MOVEMENT RECOGNITION OF FOOTBALL PLAYERS. International Journal of Industrial Engineering: Theory, Applications and Practice, 30(2). https://doi.org/10.23055/ijietap.2023.30.2.8581

Issue

Section

Work Measurement, Human Factors and Ergonomics