AUTOMATIC SUMMARIZATION OF HOCKEY VIDEOS

Video summarization is a useful technique in present world to view all the important aspects of a lengthy video. Video summarization extracts important representative shots from the video and it conveys all the semantics of the entire video in a short span of time. In sports, it plays a very important role to generate highlight of the game video spanning over many hours. Hence we propose an algorithm for automatic summarization of lengthy hockey game videos. In this method, different efficient algorithms are devised to find shot detection, penalty corner and penalty stroke detection, Umpire detection and foul detection. The method also detects all the replay shots along with logo shots. The method finally combines all the important events detected to form a summarized video. It also allows the user to create customized video summary containing user preferred events. The method is tested with many international Hockey match videos and the average efficiency is found to be 0.88.