Image Retrieval Based On The Color Textons Descriptor

In perceptually uniform color space, color differences can be measured in a way close to the human perception of two colors. A new feature descriptor, namely the color textons descriptor, is proposed in this paper according to the attribute of perceptually uniform color space. This descriptor can represent features via color differences in textons component images. It converts color image from ROB color space to CIE Lab color space, and then quantized the color images into 64 colors and find out the textons location in Lab color components. Finally, color differences are used to represent features and for image retrieval. Image retrieval experimental results have shown that the color textons descriptor has the discrimination power of color and texture features, and the performance is better than that of the edge histogram descriptor and LBP method significantly.