Abstract

The exploitation of video data requires to extract informa- tion at a rather semantic level, and then, methods able to infer “con- cepts” from low-level video features. We adopt a statistical approach and we focus on motion information. Because of the diversity of dy- namic video content (even for a given type of events), we have to design appropriate motion models and learn them from videos. We have de- fined original and parsimonious probabilistic motion models, both for the dominant image motion (camera motion) and the residual image motion (scene motion). These models are learnt off-line. Motion mea- surements include affine motion models to capture the camera motion, and local motion features for scene motion. The two-step event detection scheme consists in pre-selecting the video segments of potential interest, and then in recognizing the specified events among the pre-selected seg- ments, the recognition being stated as a classification problem. We report accurate results on several sports videos.