[4Yin-A] Poster 6

[4Yin-A-25] An AI-Based Golf Swing Evaluation Method Using Pose Estimation

〇Junwei Fan¹, Aya Yoshizawa¹, Etsuko Ichida¹, Kousuke Aoki¹ (1. INTEC Inc.)

Keywords:

Golf Swing Evaluation,Pose Estimation,Image Processing,Motion Analysis,Sports Tech

In this study, we propose a golf swing evaluation method that combines pose estimation and image processing techniques.The proposed method extracts joint point coordinate data of a user through pose estimation and automatically detects the swing initiation timing. The user’s swing is then temporally aligned and compared with a coach’s reference swing video. Swing performance is evaluated based on the movements of full-body joint points, including displacement and movement angles.Furthermore, focusing on key moments in the swing, evaluation metrics such as arm angles, arm-torso separation, and center-of-gravity displacement are defined. These metrics are visualized, and an overall swing evaluation score is calculated and presented to the user, along with feedback messages and swing evaluation comments indicating areas for improvement.Compared with existing related studies, the proposed method requires lower computational cost, making it suitable for practical deployment in real-world environments.