Event Detection Based Approach for Soccer Video Summarization Using Machine learning

Many soccer fans prefer to watch a summary of football games as watching a whole soccer match needs a lot of time. Traditionally, soccer videos were analyzed manually, however this costs valuable time. Therefore, it is necessary to have a tool for doing the video analysis and summarization job automatically. Automatic soccer video summarization is about extracting important events from soccer matches in order to produce general summaries for the most important moments in which soccer viewers may be interested. This paper presents a machine learning (ML) based event detection and summarization system for emphasizing important events during soccer matches. The proposed system firstly segments the whole video stream into small video shots, then it classifies the resulted shots into different shot-type classes. Afterwards, the system applies two machine learning algorithms, namely; support vector machine (SVM) and artificial neural network (ANN), for emphasizing important segments with logo appearance with addition to detecting the caption region providing information about the score of the game. Subsequently, the system detects vertical goal posts and goal net. Finally, the most important events during the match are highlighted in the resulted soccer video summary. Experiments on real soccer videos demonstrate encouraging results. The proposed approach greatly reduces workload and enhances the accuracy of summarizing soccer video matches with reference to both recall and precision performance measurement criteria.