Research on the analysis and optimization strategies of low-carbon sports behavior patterns based on deep learning
Abstrak
This study proposes a method for the analysis and optimization of low-carbon sports behavior patterns based on deep learning. Combining a sports behavior recognition algorithm with implicit label supervision and a module for extracting key individual characteristics, the method effectively integrates spatial and temporal features, significantly enhancing the ability to recognize complex movement patterns. This study addresses the critical need for sustainable sports analytics that significantly advances low-carbon behavior recognition in professional basketball. Training and testing on the NCAA and UCF-BBall dataset shows that the proposed model achieves 89.3 % and 85.7 % accuracy in sports behavior identication respectively. The proposed method excels in identifying various movement patterns in basketball games, validating its effectiveness in recognizing sports behavior patterns, improving athletic efficiency, and reducing unnecessary energy consumption.
Topik & Kata Kunci
Penulis (1)
Long Chu
Akses Cepat
- Tahun Terbit
- 2025
- Sumber Database
- DOAJ
- DOI
- 10.1016/j.sasc.2025.200387
- Akses
- Open Access ✓