Fractal technique for color texture segmentation

In this paper, the problem of texture segmentation for color images is discussed. To achieve an effective representation of color texture features, color information of the image is used. By analyzing both the volume between color surfaces of image pixels and the single color components, corresponding fractal features are acquired. We then proceed to apply median filtering and an unsupervised clustering algorithm to complete the segmentation on the feature image. Experiment results show that the use of fractal techniques is an efficient approach to color texture segmentation.