Multimodal Semantics Extraction from User-Generated Videos

Figure 4

View from the top of a public happening (sport) with elliptical layout (stadium).