AN NOVEL SHORT TRACK SPEED SKATING TRACKING SYSTEM
In this project, we propos a novel and effective tracking system to track high-speed skaters. A novel registration method is employed to automatically discover key frames to build the panorama. Then the homography between a frame and the real world rink can be generated accordingly. Aimed at several challenging tracking problems of short track skating, a novel multiple objects tracking approach is proposed which includes: Gaussian Mixture Models (GMMs), evolving templates, constrained dynamical model, fuzzy model, multiple objects template initialization and evolution, etc. The outputs of the system include spatial-temporal trajectories, speed curves, 2D-reconstruction animations. All the information is invaluable for sports experts. Experimental results demonstrate the effectiveness and robustness of the proposed system, and the proposed approach can be applied to many areas.
Original Videos
Two testing videos are obtained from 2005/2006 BOSIDEN ISU (International Skating Union) world cup short track speed skating. One is Men 500m finals which has 1071 frames with frame rate 25 frames/second. The other one is Women 500m semi-finals which has 1156 frames with the same frame rate as video 1.
Tracking Results
Reonstruction Results
Sync Results
Comparions
For more information, please contact with Yuxuan Wang
Email: yuxuan.wang@aggiemail.usu.edu