A novel approach for contents-based e-catalogue image retrieval based on a differential color edge model

Abstract

In this paper, we propose a new color edge model and color edge histogram descriptor for contents-based image retrieval. The edge descriptor proposed by MPEG-7 standard is a representative approach for the contentsbased image retrieval using the edge histogram that is derived from a gray-level image. This paper introduces a novel method that extracts color edge information from spectral color images rather than monochrome images and a new color edge histogram descriptor for the contents-based image retrieval. The proposed color edge model is obtained in two phases. In the first phase, we characterize the R,G,B channel components as a linear map (the differential) and impose a statistical interpretation on this map by considering the behavior of the map as it applied to unit normed vectors in the second phase. As a result, the constructed edge model will be expressed in a statistical fashion and will provide a mechanism to determine the possibility of the edge existence. The color edge histogram based on the direction of the color edge model is subsequently applied to the contents-based e-catalogue image retrieval. For the evaluation, the results of image retrieval using the proposed method are compared with those of image retrieval using the edge descriptor by MPEG-7 and other approaches. The experimental result supports the efficiency of the proposed method. © Springer-Verlag 2004.