A New Framework for Goal Detection Based on Semantic Events Detection in Soccer Video

Soccer video processing and analysis to find criticalevents such as occurrences of goal event have been one of theimportant issues and topics of active researches in recent years. Inthis paper, a new role-based framework is proposed for goal eventdetection in which the semantic structure of soccer game is used.Usually after a goal scene, the audiences’ and reporters’ soundintensity is increased, ball is sent back to the center and the cameramay: zoom on Player, show audiences’ delighting, repeat the goalscene or display a combination of them. Thus, the occurrence ofgoal event will be detectable by analysis of sequences of aboveroles. The proposed framework in this paper consists of four mainprocedures: 1- detection of game’s critical events by using audiochannel, 2- detection of shot boundary and shots classification, 3-selection of candidate events according to the type of shot andexistence of goalmouth in the shot, 4- detection of restarting thegame from the center of the field. A new method for shotclassification is also presented in this framework. Finally, byapplying the proposed method it was shown that the goal eventsdetection has a good accuracy and the percentage of detectionfailure is also very low.

کلیدواژه ها:

Soccer Video Processing ، Event Detection ، Shot Boundary Detection ، Shot Classification ، Field Extraction ، Field Center Detection