Fuzzy c-means (FCM) clustering technique has been widely applied in image segmentation. However, it is quite sensitive to the various noises or outliers. In order to further improve the segmentation accuracy and robustness to noise, a fuzzy nonlinear weighted local information c-means clustering method is proposed for unsupervised segmentation of noisy images in this paper. First, a fuzzy nonlinear weighted factor including both the spatial distance of local window and its gray-level difference

more »

... in the similarity measure is introduced to guarantee noise insensitiveness. Second, the spatial neighbor constraints are also taken in the membership function to enhance fuzzy clustering performance. The performance of this algorithm is evaluated by two images: synthetic images and brain MR images, and the experimental results demonstrate that the proposed algorithm is more robust to noise and effectively preserves the image detail than FCM algorithm and its variants.