Yen-Hua Chen

Tri-Service General Hospital, National Defense Medical University

Abstract Title: An AI-Assisted Framework for Customized Movement Assessment and Training: Enhancing Rehabilitation Outcomes and Clinical Education Efficiency

Biography:

Research Interest:

Background: Efficient and accurate movement assessment is essential in physiotherapy but often depends on clinician experience, leading to variability in clinical decision-making. We developed an AI-assisted framework (eAi-Agent) to support accurate movement assessment and training. Methods: Developed by EzAi Health Co since 2017 and clinically implemented since 2025, this framework is structured around the REPAIR model: Recognize pain, Evaluate weakness, Physical modeling, Analyze, Issue correction, and Reinforce control. Utilizing camera-based video analysis, the system assesses posture, joint mobility, muscle activation, and alignment to generate a precise functional diagnosis. Personalized training is then designed following the RSAI principle—Release, Stretch, Activate, Integrate—and supported by home practice videos and remote feedback from physiotherapists. The system incorporates muscle imbalance theories and identifies compensatory patterns. AI algorithms provide automated interpretation and generate individualized training recommendations. Results: This framework enables rapid and objective assessment of movement quality and postural control. Static posture can be captured through a 3-second image acquisition, while functional movements are recorded via 6-second video clips. Within approximately 10 seconds after data upload, the system generates a comprehensive and precise assessment report. The AI model identifies dysfunctional movement patterns and analyzes underlying muscle imbalances, highlighting specific muscle groups associated with impaired control. Based on these findings, the system provides personalized, targeted exercise prescriptions, including corrective movement strategies and muscle strengthening programs. The framework also translates complex biomechanical concepts into structured and interpretable outputs, facilitating training and learning efficiently. Conclusion: The eAi-Agent offers a rapid and precise tool to detect movement dysfunctions and guide individualized interventions, supporting functional independence across home, clinical, and community environments. Keywords: eAi-Agent, movement assessment, customized and remote training