A graph-theoretical clustering based anchorperson shot detection for news video indexing

Anchorperson shot detection is of significance for video shot semantic parsing and indexing information and clues extraction in content-based news video indexing and retrieval system. This paper presents a model-free anchorperson shot detection scheme based on the similarity among the anchorperson key frames throughout each news program. First, a news video is segmented into video shots with any effective video syntactic parsing algorithm. For each shot, a frame is extracted from the frame sequence as a representative key frame. Then the graph-theoretical clustering algorithm is performed on the key frames to identify the anchorperson frames. The anchorperson shots are further distinguished from other news video shots. The proposed scheme achieves a precision of 100% and a recall of over 97.69% in the anchorperson shot detection experiment of 217 news stories.