An OpenPose-Based Exercise and Performance Learning Assistant Design for Self-Practice Yoga
Cheng-Hsien Lin, Shih-Wei Shen, Irin Tri Anggraini, Nobuo Funabiki, Chih‐Peng Fan · 2021 IEEE 10th Global Conference on Consumer Electronics (GCCE) · 2021
In this paper, a performance learning assistant design based on OpenPose is studied for self-practice Yoga. Firstly, the skeleton information of the human body is extracted through OpenPose. Secondly, to calculate the angle values of the selected keypoints, the vectors of the user and the instructor are obtained based on the center points. Next, by using the angle values, the angle differences between the user and the instructor are calculated. Finally, the scoring system is developed to calculate the total scores for evaluation of the user’s Yoga posture. The experimental results show that the proposed design effectively detects the posture differences between the user and the instructor. Moreover, the proposed design performs up to 5.5 frames per second (FPS) on the GPU-based embedded platform.