Multimodal Semantics Extraction from User-Generated Videos

Figure 9

Detection of in-view cameras (view from the top).