A robust video searching method for a video signal query captured by a portable camera

摘要

We propose a search method for detecting a query video signal fragment in long video recordings. The query signal is assumed to be captured by a portable camera in the real world. A major problem in this kind of search is that the features of the query video may include distortions due to terminal characteristics or illumination distortion. The method proposed here comprises local time-space-region normalization and robust subspace spanning. The former is used to make features invariant to terminal characteristics or illumination distortion, and the latter to choose colors and spatial area that minimize the effect of feature distortions. Experiments using cameras in the real world show the proposed method is effective.