Texture-Based Image Retrieval by Edge Detection Matching GLCM

Abstract

Content-Based Image Retrieval (CBIR) system is emerging as an important research area, users can search and retrieve images based on their properties such as shape, color and texture from the image database. Usually texture-based image retrieval just consider an original image of coarseness, contrast and roughness, actually there is much texture information in the edge image. This paper proposed a novel approach to retrieve images by texture characterization using a composition of edge information and co-occurrence matrix properties. The proposed method gives encouraging results when comparing its retrieval performance to that of the traditional co-occurrence matrices and Yaopsilas approach.