Color Texture Image Segmentation Based on Neutrosophic Set and Nonsubsampled Contourlet Transformation

In this paper, an automatic approach for image segmentation based on neutrosophic set and nonsubsampled contourlet transformation for natural images is proposed. This method uses both color and texture features for segmentation. Input image is transformed into LUV color model for extracting the color features. Texture features are extracted from the grayscale image. Image is then transformed into Neutrosophic domain. Finally, image segmentation is performed using Fuzzy C-means clustering. Clusters are adaptively calculated based on a cluster validity analysis. This method is tested in natural image database. The result analysis shows that the proposed method automatically segments image better than traditional methods.