Summary: Players and the ball in soccer video are supposed to inside the grass play field. A shape analysis-based approach is put forward to identify the players and the ball from the roughly extracted foregrounds which are obtained by a learned color histogram-based playfield detector and connected component analysis. Euclidean distance transform is employed to extract skeletons for every foreground blobs, and then skeleton analysis is performed to remove false alarms (non-players and non-ball) and cut-off the artifacts (mostly due to playfield lines) based on skeleton pruning and reverse Euclidean distance transform. Details are given in paper MCAM'07.
