A Low Cost Algorithm for Expected Goal Events Detection in Broadcast Soccer Video

Recently, mining information in sports video data, especially soccer video, has become an active research topic. In this paper, a new algorithm for detection of expected goal events in soccer video is proposed. The proposed algorithm is composed of two main steps. Firstly, video is segmented into its constituent shots and these shots are categorized into two groups, namely long shots and non-long shots. Secondly, long shots are examined to detect expected goal events. Our scheme uses the playfield boundary feature for shot boundary detection, shot classification, and expected goal events detection. Therefore, this feature is extracted only one time and the algorithm is speedy. It is noticeable that the proposed algorithm is robust to blurring and spatial downsampling. Moreover, experimental results on various broadcast soccer videos show that our algorithm can achieve high accuracy and especially high speed.