Color texture segmentation using feature distributions

Abstract

A new method for color textured segmentation using feature distributions is proposed in this study. First, the distributions of color and local edge patterns are used to derive a homogeneity measure for color textured regions. Then, a coarse-to-fine method based on the homogeneity measure is employed to achieve the goal of color texture segmentation. The proposed method unifies color and edge features to solve the color texture segmentation problem rather than simply extend gray-level texture analysis to color images, or analyze only spatial interaction of colors in a neighborhood. In addition, the proposed method is simple but effective. The execution time is fast, the error rate for collages of real texture is low and the segmentation results for natural scenes are visually satisfactory. Finally, no a priori knowledge about the number and types of textures or the number of regions are required in the proposed method. The feasibility and effectiveness of the proposed method have been demonstrated by various experiments.