Scholarlink

<p>Background: Home-based fitness training requires automated systems for exercise quality assessment and real-time posture correction without professional supervision. Purpose: We develop an AI-powered virtual fitness system using spatiotemporal skeleton analysis for automatic exercise quality assessment. Methods: We propose Anatomical-Prior Sparse Attention (APSA), integrating biomechanical constraints into graph neural networks. APSA hierarchically models spatial and temporal dependencies through anatomically-informed attention, focusing on kinematically meaningful joint interactions while pruning implausible connections. The system processes 3D skeletal keypoints from standard RGB cameras. Results: Evaluation on two public datasets (IntelliRehabDS and Kimore) shows APSA achieves 76.9% accuracy in quality classification and 0.789 F1-score in error detection, outperforming recent 2024 methods by 1.1–8.5 percentage points while maintaining real-time performance (8.2&#xa0;ms per sequence). The system identifies common errors including knee valgus (84.3% accuracy) and excessive trunk lean. Conclusions: The biomechanically-informed attention mechanism enhances accuracy and interpretability, enabling scalable deployment in home-based training and remote rehabilitation.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Spatiotemporal skeleton based learning for automatic exercise quality assessment and real time posture correction

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年5月12日

查看全文
  • Xinyinan Wang

摘要

Background: Home-based fitness training requires automated systems for exercise quality assessment and real-time posture correction without professional supervision. Purpose: We develop an AI-powered virtual fitness system using spatiotemporal skeleton analysis for automatic exercise quality assessment. Methods: We propose Anatomical-Prior Sparse Attention (APSA), integrating biomechanical constraints into graph neural networks. APSA hierarchically models spatial and temporal dependencies through anatomically-informed attention, focusing on kinematically meaningful joint interactions while pruning implausible connections. The system processes 3D skeletal keypoints from standard RGB cameras. Results: Evaluation on two public datasets (IntelliRehabDS and Kimore) shows APSA achieves 76.9% accuracy in quality classification and 0.789 F1-score in error detection, outperforming recent 2024 methods by 1.1–8.5 percentage points while maintaining real-time performance (8.2 ms per sequence). The system identifies common errors including knee valgus (84.3% accuracy) and excessive trunk lean. Conclusions: The biomechanically-informed attention mechanism enhances accuracy and interpretability, enabling scalable deployment in home-based training and remote rehabilitation.