CBIR using Texels of colour Fuzzy Textons

Abstract

This paper presents an innovative feature representation method for image retrieval, using a fuzzy texton approach. Several image retrieval techniques represented in researches have proved only up to certain level. The texton based approach which highlights discrimination of the low level features like color, texture and shape, are elegant than many other peer content based image retrieval (CBIR) techniques. The Texels of Fuzzy Texton approach (TFT), proposed in this paper, and primarily it retrieve images efficiently without much increasing the computational costs. The proposed technique involves HSV (Hue, Saturation, and Value) color model instead of RGB color model, as the former model was illumination invariant. With the involvement of HSV color model, the proposed TFT approach produced better recall. The test results dominated in efficiency than all the predecessor peer techniques like Texton Co-occurrence Matrix (TCM), Texton Multi Histogram (TMH) and Texels of RGB Texton (TT).