Compact and efficient CBCD scheme based on integrated color features

Abstract

Content-Based Copy Detection (CBCD) is an emergent research topic and has been extensively studied recently. Color is one of the most important and easily recognizable features of visual content. To address efficiency and effectiveness issues of CBCD systems, in this article a compact and computationally inexpensive CBCD scheme using MPEG-7 Dominant Color Descriptor is proposed. This paper proposes a novel dominant color extraction technique, which solves the intrinsic problems of existing color clustering techniques. Experimental results show that the proposed scheme improves detection accuracy up to 38%, and also supports significant reduction in total computational cost up to the extent of 91%, when compared with that of existing schemes against various video transformations. © 2011 IEEE.