A novel approach to detect fours in cricket videos

This paper proposes a novel approach to detect or identify fours in cricket videos. The problem in automatic four detection arises because ball detection and boundary detection is not a very easy task. Problems in ball detection arises because the ball could get occluded by the players, the speed of the ball usually very high and due to the high speed of the ball, its shape and size may get distorted, several cameras captures one complete shot. Problem in boundary detection arises as there can be multiple straight lines in the image so it is hard to choose the correct boundary line. The Proposed algorithm consists of three steps. We identify the boundary in the video frames. Next we try to identify ball in the video frames in which the boundary has been detected. Once both are detected in a frame, we compute the distance between the ball and the boundary. If the distance is less than the threshold then there will be four else not. The proposed approach provides satisfactory results for identifying four in the cricket. Fifteen video shots from cricket are used to evaluate our algorithm.