OpenPose Based Yoga Poses Difficulty Estimation for Dynamic and Static Yoga Exercises
Wan-Chia Huang, Cheng-Liang Shih, Irin Tri Anggraini, Yanqi Xiao, Nobuo Funabiki, Chih‐Peng Fan · 2023
Assessing the difficulty of a Yoga pose is a challenging task, often dependent on individuals’ subjective judgement or consultation with professionals to choose poses that fit their needs. In this study, a preliminary difficulty calculation method for estimating the difficulty of Yoga poses is proposed, and it is applied to the Yoga poses in the proposed system in order to provide a user-friendly difficulty reference. OpenPose is applied in the system to detect user’s movements and extract the key-points information. In the experimental results, seven users are inquired to participate in practicing Yoga by the proposed system, and they are asked to provide the difficulty levels rating from 1 to 5. These subjective difficulty levels are used to compare the estimated difficulty levels. The results indicate a positive correlation between the subjective and estimated difficulty results.