EE 368 Project : Football Video Registration and Player Detection

This paper focuses on methods that can be used in football analysis by showing our implementation and results of automatic football video registration and player detection. Some of the first methods we implemented like SIFT Feature detection for video registration and background detection for player detection were not successful. However, we implemented more integrated methods like line mapping using hough transform and unique feature identification for video registration and optical flow and color filtering for player detection which were more successful but could still be improved.