Development of a Device of Table Officials using OpenPose in Basketball Competitions

Abstract

We propose a new device that supports table officials in basketball games with using OpenPose. OpenPose uses deep learning in image recognition, and visualizes human movement by detecting human body, hand, and facial keypoints on a single image. In official games of the basketball, table officials are necessary for the progress of games in addition to referees. Table officials mainly consists of a scorer, an assistant scorer, a timer and a shot clock operator. The proposed device of the table officials can read the information such as a score and individual personal fouls by analyzing gestures of referees. Furthermore, the device can display and record the information. We think that the device can reduce some mistakes of table officials.