Multimodal Semantics Extraction from User-Generated Videos

Figure 3

View from the top of a public event (live musical performance) with Nonelliptical layout (nonstadium).