A Novel Shot Detection Algorithm Based on Information Theory

As a crucial step in content-based video retrieval, shot boundary detection has attracted much more research interesting in recent years, but many of the algorithms proposed in the literatures are not robust to mosaic, flash lamp and motion of camera and object. A novel shot boundary detection algorithm based on information theory is presented in this paper. It can detect cut and gradual shot simultaneously and distinguish them. The algorithm is not intensive to all of those interference factors. At first the features of color and texture are extracted by wavelet transform, then the dissimilarity between two successive frames are defined which colligates the mutual information of color feature and the co-occurrence mutual information of texture feature. The threshold is adjusted adaptive based on the entropy of the continuous frames, and it is not depend on the type of video and the kind of shot. The efficiency of the proposed algorithm is extensively tested on news, football and other videos, and 91.9% recall with 90.8% precision has been achieved.