Abstract
The purpose of this article is to analyze current vision-based systems from a soccer video semantic point of view such as video summarization, features analysis, and provision of augmented information. Currently, computer vision techniques are applicable in a challenging soccer context. Scene interpretation is performed based on the complexity of the semantic. For each area of vision-based systems, computer vision methodologies are analyzed along with their strengths and weaknesses. We have also investigated whether the existing approaches are equally applicable for real-time soccer video semantic analysis.