Abnormal Motion Detection Based on Derivative Dynamic Time Warping Using Motion Data of Bowling-Pitching Posture
Eikou Taira, Genki Suzuki, Hiroyuki Shioya · 2023
Bowling has many elderly users who need to acquire the correct throwing posture to prevent injury. This paper presents a method to detect abnormal motion in the bowling throwing posture using anomaly detection. The distance is calculated at each data point by inputting the coordinate data of each body part obtained from motion capture into Derivative Dynamic Time Warping. Then, anomaly detection using the Hotelling theory is used to detect the movement of the anomalous part when the ball is thrown. The proposed method shows better results than other comparison methods of abnormality detection.