A Study on Sports Player Tracking based on Video using Deep Learning
Jung-Soo Lee, Sung-Won Moon, Do-Won Nam, Jiwon Lee, Ah Reum Oh, Wonyoung Yoo
Abstract
Tracking the players in the game is essential for the correct evaluation of the players. In order to obtain evaluation indicators for players, such as the moving distance and average speed during a game, it is necessary to continuously track the location and trajectory of the player. The players' movement and events' analysis in the games are mainly recorded by professional analysts. To compensate for this, some sports fields use image processing tools for player tracking and event analysis. In this paper, we study the method that shows excellent performance in the detection and tracking objects recently using deep learning, and discuss how to apply it to sports field.