Modified Color Texton Histogram for Image Retrieval

Abstract

In this paper, HSV based text on histogram (HSV-TH) is proposed for content based image retrieval (CBIR). The HSV-MT is proposed in contrast to the RGB based text on histogram (RGB-TH). The proposed HSV-TH method is based on Julesz's textons theory, and it works directly on nature images as shape descriptor and a color texture descriptor. HSV-TH integrates the advantages of co-occurrence matrix and histogram by representing the attribute of co-occurrence matrix using histogram. The retrieval results of the proposed method are tested on the different image databases i.e. Corel 1000 (DB1), Corel-10,000 (DB2) and MIT VisTex(DB3). Two experiments have been carried out for proving the worth of our algorithm. The results after being investigate show that a significant improvement in terms of their evaluation measures as compared to RGB-TH.