深層学習による三次元姿勢推定を用いたアーチェリーの射型教示システム

抄録

Archery is a sport in which it is important to repeat the same shooting form and to improve its repeatability. For this reason, it is essential to understand and master the basic shooting form. However, there is a shortage of instructors, and in university club activities, the burden of teaching is placed on the senior students. A possible solution to this problem is the use of a practice support system based on sports video analysis, but existing systems have a fixed positional relationship between the camera and the person to be measured, and it is necessary to create the same environment each time the system is used. Therefore, this study proposes a sports practice support system that uses 3D skeletal point information extracted from video images to evaluate the learner’s shooting style in terms of points, eliminating the need to fix the positional relationship between the camera and the learner and allowing easy recognition of points that need improvement.