Modeling visual information by spatio-temporal patterns to analyze event tactic in sports video

Although event detection in sports videos has been gaining some remarkable results, tools for exploring and exploiting these event clips on their own demands are far from user's expectation. In this paper, a novel method using spatio-temporal patterns to model visual information tailored to analyze event tactics in sports videos without tracking objects is introduced. The major contribution of the proposed method is to represent visual information by using a sequence of symbols. Each symbol represents a distribution of players in a certain period of time. By using these sequences, event tactics are detected, explained, and visualized to help users understand event tactics without watching whole clips. Moreover, users could query by an example or by a text to find all events sharing the same tactic. Thorough testing with over 100 goal clips in soccer domain demonstrates the superiority of the proposed method in terms of precision recall ratios.