Contribution of a color classification to soccer players tracking with snakes

In this paper, we present a soccer players tracking method by successive color images analysis. Each player who is present in the images is modelized by a snake. The snake is computed after a color classification of the pixels in order to recognize the team of each player. The classification process uses a supervised learning scheme based on the colorimetric analysis of the players suits. So, each player is tracked even if he is close to a player of the other team.