Human’s Reaction Time Based Score Calculation of Self-practice Dynamic Yoga System for User’s Feedback by OpenPose and Fuzzy Rules

In this study, the proposed self-practice dynamic Yoga system at home with instructor-user synchronization provided a new score calculation as user’s feedback. By following the guide of the instructor’s video, a proper score feedback indicates whether the user’s dynamic Yoga is good or bad, and helps users to understand how to improve their movements and postures. In the new score calculation, it considers two types of scores. The first type of scores is "Shape_scores", and the system compared the similarity of static postures between the instructor and user, and then the "Shape_scores" is calculated through the scoring system based on fuzzy rules with the angle difference values between the instructor and user. The second type of scores is "Delay_scores" by appending on the human’s average reaction time and average recognition time as thresholds to set the five different regions to give the scores for delays. By weighting these two scores and then summing them up, a new score is calculated finally. In the experimental results, the score schemes between the "Shape_scores" and the "Delay_scores" are discussed, and the tester is a lazy mode user who means the user simulates the elderly to observe the influence on the two different scores calculations.