Related Experiment Video
Updated: Jun 11, 2025
12:39
A Methodology for Capturing Joint Visual Attention Using Mobile Eye-Trackers
Published on: January 18, 2020
7.6K
Multistream Adaptive Attention-Enhanced Graph Convolutional Networks for Youth Fencing Footwork Training
Yongjun Ren1, Huinan Sang1, Shitao Huang1
1School of Computer Science, Nanjing University of Information Science and Technology, Nanjing, JS,China.
Pediatric Exercise Science
|October 1, 2024
Summary
Artificial intelligence (AI) enhances adolescent sports training by accurately recognizing fencing footwork, reducing ineffective exercises and training burden. This AI-driven approach personalizes movement corrections for improved athletic development.
Area of Science:
- Sports Science
- Artificial Intelligence
- Biomechanical Analysis
Background:
- Adolescent athletes face excessive training burdens due to intense sports like fencing.
- Traditional training methods lack personalized feedback and timely movement corrections.
- Ineffective exercises contribute to physical harm and hinder athletic development in young athletes.
Purpose of the Study:
- To develop an AI-driven action recognition algorithm for adolescent athletes.
- To reduce ineffective exercises and alleviate the training burden in sports.
- To provide timely movement corrections and personalized training plans.
Main Methods:
- Proposed an action recognition algorithm tailored for adolescent athletes.
- Utilized multimodal input data with shared network structures, attention mechanisms, and adaptive graphs.
- Employed a multibranch feature fusion method for final action classification.
Main Results:
- Achieved 93.3% accuracy, 95.8% precision, and 94.5% F1-Score on the fencing footwork dataset 2.0.
- Effectively recognized actions across varying adolescent heights and speeds.
- Demonstrated superior performance compared to traditional training assessment methods.
Conclusions:
- AI-based solutions significantly improve training efficiency for adolescent athletes.
- The developed algorithm effectively reduces the overall training burden.
- Personalized, AI-driven feedback enhances athletic performance and safety.
