Development and Training Strategy of Badminton Action Recognition System Under the Background of Artificial Intelligence
To improve the intelligence level of badminton training and match analysis, this study discusses a badminton action recognition system based on deep learning. It optimizes spatiotemporal feature extraction, multimodal data fusion, and computational efficiency to address current issues such as insuff...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2025-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/10973125/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | To improve the intelligence level of badminton training and match analysis, this study discusses a badminton action recognition system based on deep learning. It optimizes spatiotemporal feature extraction, multimodal data fusion, and computational efficiency to address current issues such as insufficient recognition accuracy, high computational resource consumption, and limited real-time performance. In the experimental section, the performance of three mainstream models—Spatial-Temporal Graph Convolutional Network (ST-GCN), Vision-Attention Transformer for Real-time Motion Recognition (VATRM), and Multi-Modal Network for Sports Action Recognition (MM-Net)—is compared from two dimensions: recognition performance and computational efficiency. The experimental results show that the optimized system achieves accuracies of 0.943, 0.967, and 0.912, higher than other comparison models. Additionally, in terms of F1 score, the optimized system scores 0.902 in the defense and transition action groups, significantly outperforming MM-Net (0.864) and ST-GCN (0.878). In terms of computational efficiency, the inference times of the optimized system are 12.4ms, 10.8ms, and 13.1ms, faster than ST-GCN and MM-Net, with memory usage as low as 487MB, 452MB, and 501MB, making it suitable for real-time mobile applications. Therefore, this study contributes to intelligent sports training, sports data analysis, and badminton technique optimization. |
|---|---|
| ISSN: | 2169-3536 |