Semantic analysis of soccer video based on a fuzzy event mining approach

This paper presents an approach for event detection in broadcast soccer video. The goal of this paper is to propose a flexible system being able to tackle different situations that may occur in video sequences. To achieve this goal, we design a fuzzy rule-based reasoning system which adopts statistical information from a set of audiovisual features that are organized in a hierarchical structure as its input and produces semantic concepts corresponding to the occurred events. A set of tuples is created using information derived from training set of feature vectors and their related events. We extract the hidden knowledge among the tuples by constructing a decision tree (DT). A set of rules is generated by traversing each path from root toward leaf nodes of constructed DT. These rules are inserted in rule base of designed fuzzy system and employed by fuzzy inference engine to do decision-making process and predict the events. Experimental results conducted on a large set of soccer videos demonstrate the effectiveness of the proposed approach.