Event Detection and Recognition Using HMM with Whistle Sounds
Hiroki Itoh, Tetsuya Takiguchi, Yasuo Ariki · 2013
In this paper, we propose a new method to detect and recognize events robustly in a soccer game. Based on the players density and speed, the events are detected and recognized using Hidden Markov Model (HMM). However, it is difficult to detect "free kick" and "throw in" because these events occur anytime and anywhere. In a soccer game, some event occurs when the referee blows a whistle or a ball is out of field. Therefore, we improve the detection accuracy of the events such as "free kick" and "throw in" by using these information when they occur. Also, event recognition is performed by an integration method of the results obtained using two types of HMMs: one is for players and the other is for a ball.