Analysis and Feedback of Baseball Pitching Form with use of Kinect

Abstract

This paper introduces an automatic analysis and feedback system for amateur baseball pitching learners with use of Microsoft Kinect. Compared with conventional 3 -D motion capture systems, Kinect has advantages of reasonable cost and easiness of application system development for physical exercises. The authors developed a similar system for flying disc throw, which is disclosed in Tamura, Yamaoka et al. (2013) and Tamura, Uehara et al. (2013). In this paper a target motion is moved to baseball pitching, so focused body parts and judgment criteria were changed, although utilizing the same Kinect platform. The proposing system acquires postures and motions of amateur baseball pitchers, and judges them in 2 criteria: (1) maximum angle of elbow, and (2) hand twist, which were decided by a training expert. It also displays feedback messages to improve their motions. As a result, novice testees of the target group showed significant improvement of their pitching motions.