Multimodal feature extraction and fusion for semantic mining of soccer video: a survey

摘要

This paper presents a classified review of soccer video analysis works. The existing approaches in the aspects of highlight event detection, video summarization and retrieval based on video stream, ball and player tracking for provision of match statistics, technical and tactical analysis and application of different sources in soccer video analysis have been surveyed. In addition, some major existing commercial softwares developed for video analysis are introduced and compared. With regard to the existing challenge for automatic and realtime provision of video analysis, different computer vision approaches are discussed and compared. Audio, video and text feature extraction methods have been investigated and the future trends for improvement of the reviewed systems have been introduced in terms of response time optimization, increase of precision and eliminating the need of human intervention for video analysis.