Depth Camera-based Monitoring and Simulation System for Weight Training Assessment

This paper introduces a depth camera-based monitoring and simulation system, specifically designed for assessing weight training. Recognizing the limitations of Internet of Thing (IoT) devices and wearable sensors in providing direct exercise guidance, this study employs a depth camera for real-time posture monitoring. The system integrates skeleton recognition to capture joint trajectories, feeding them into an OpenSim-based upper limb model for detailed muscle and joint analysis. This approach allows assessment of muscle fiber and tendon forces through inverse kinematics and dynamics calculations. Camera setup tests in various exercise scenarios ensure optimal tracking accuracy. The findings indicate effectiveness of the system in enhancing efficiency in weight training, which provides a method for non-contact training monitoring.