BROWSE
Detail
Conference Paper
Multiple Player Tracking in Indoor Sports with Multi-cameras
Cited - time in 
Share 
Share
- Authors
- Yookyung Kim, Gi-Mun Um, Kwang-Yong Kim, Kee-Seong Cho, Hyungmin Kim, Jong-Il Park
- Issue Date
- 2016-01
- Citation
- International Workshop on Advanced Image Technology (IWAIT) 2016, pp.1-2
- Language
- English
- Type
- Conference Paper
- Abstract
- Automated system for analyzing a sport game is very useful. This paper proposes the system of tracking multiple players with multi-cameras under indoor sport environments, especially for basketball games, and visualizing their locations on the screen that displays the entire court area. The proposed system consists of four steps: pre-processing, unifying a coordinate system among multi-cameras, tracking multiple players, and representing results under a unified coordinate system. The first step eliminates the effects of lens distortions. In the second step, multi-camera images that cover different court areas are projected onto the common court area using homography estimation and shared the coordinate system. The third step of tracking utilizes particle filter, which is adequately used for indoor sports that often involve abrupt motion changes and player-to-player occlusion, instead of using a simple motion model. The last step shows the player’s footages on the entire court map throughout the game. This automated multiple player tracking system, which is carefully designed for indoor sports, was validated with real game scenarios while demonstrating robust performance even with frequent motion changes.
- KSP Keywords
- Multi-camera, Multiple player tracking, Particle filters, Pre-processing, Robust performance, Second step, Tracking System, Unified coordinate system, abrupt motion, automated system, game scenarios