3D Swing Trajectory of the Club Head from Depth Shadow

Abstract

Using depth sensors like Microsoft Kinect for gesture recognition has recently drawn much academic attention. It is possible to recognize a user's body parts from depth images and find joint positions. Yet, it is not easy to track the club heads in fast swing motion. The present paper draws on single depth images to propose a new method of tracking the hand and club head paths in swing motion. Although the club head depth information cannot be directly gained from a depth sensor, a 3D trajectory can be generated by tracking and refining the hand and club head positions based on the depth-related shadow region information (depth shadow), hand depth information and club head depth information in the initial swing. The present experimental findings indicate the proposed method can easily capture the 3D hand and club head positions in driver tee shot motion in a small room.Index Terms-Kinect, depth shadow, club head trajectory, depth image, hand position.