Fuzzy C-means Thresholding for a Brain MRI Image Based on Edge Detection

Fuzzy C-means thresholding with a Gaussian filter based on edge-detection technique is proposed to segment the cluster-edge-detection boundary in brain magnetic resonance imaging (MRI). With an MRI scan as an input image, fuzzy c-means (FCM) clustering is used to create image clusters, which show white and gray matter. In this process, 5% Gaussian blur noise is added to the image clusters. A Gaussian filter is used to remove or reduce the Gaussian blur noise. Canny edge-detection method was used for the edge detection of a boundary of cluster image after removing the blur noise. Then, the boundary of the cluster image is segmented and extracted by thresholding. The proposed method is successfully utilized to segment the boundary of cluster images. The method efficiently produces low mean-squared error (MSE) and peak signal-to-noise ratio (PSNR) for a boundary of cluster images compared with previously reported methods.