Similarity-based Form Visualization for Supporting Sports Instructions
Souta Akiyama, Nobuyuki Umezu · 2022 IEEE 4th Global Conference on Life Sciences and Technologies (LifeTech) · 2022
We propose a support system for instructing sports forms by comparing the forms of a user and those of a model using their similarity and visualizing differences. Captured images of the model performance are scaled to fit to those of the user and overlaid to visualize differences between them. Our system has 2 modes, A) practice mode and B) test mode. In mode A), the user repeats motions to receive real time feedback using similarity based on body joint positions. Mode B) visualizes motion trajectories of the model and user to differentiate their motions. Our system also outputs advice for user’s improvements such as adjusting motion timing. We conducted a series of experiments with an implemented prototype of the proposed method. Although our method’s superiority was not clearly confirmed, we have found several advantages in our method.