Automatic soccer archive summarization using time constraint

To manage a massive growth of sport videos, we require to summarize the content into more compact and dominant representation. The previous research projects proposed soccer summarization by using highlights or scene playing; however, they were unable to do it in a given time period. In this paper, we propose an automatic soccer video summarization that can be done within the time constraint. Our proposed approach consists of three parts: shot boundary detection, event detection, and summarization. Our proposed method evaluated by using 10 soccer matches from the FIFA World Cup 2014. Experimental results show the performance that the most users satisfy at "good" level.