A method for video shot boundary detection based on HSV color histogram and DPHA feature

Abstract

With the explosion of video data, the demand of efficient video content analysis grows rapidly. Video shot boundary detection is the first and significant step for content based video retrieval (CBVR), which is one of the hottest research highlights. The histogram method is a major algorithm which is a reasonable compromise between processing and complexity. In this paper, a new method is proposed to improve the precision and integrity of shot boundary detection (SBD). The HSV color histogram method is used to be the first step of our detection and the histogram difference between two adjacent frames is calculated. Then, the dct perceptual hash algorithm (DPHA) is utilized to make further detection, which eliminates the fake boundaries from the results gave above. Experimental results indicate the improvement in precision and recall ratio of the proposed method.