Image Retrieval System Based on Color and Texture Features

Abstract

The local binary pattern (LBP) operator and its variants extract the textural information of an image by considering the neighboring pixel values. A single or joined histogram can be derived from the LBP code which can be used as an image feature descriptor in some applications. However, the LBP-based feature is not a good candidate in capturing the color information of an image, making it less suitable for measuring the similarity of color images with rich color information. To overcome this problem, we propose a fast and efficient indexing and image search system based on color and texture features. The color features are represented by combining 2D histogram and statistical moments, and texture features are represented by the local binary pattern (LBP). To assess and validate our results, many experiments were held in color space HSV. Detailed experimental analysis is carried out using precision and recall on colored Brodatz, KTH TIPS, Stex, and USPTex image databases. Experimental results show that the presented retrieval method yields about 8% better performance in precision versus recall and about 0.2 in average normalized modified retrieval rank (ANMRR) than the method using wavelet moments.