FCM with Super-Pixel Approach for Interstitial Lung Disease Image Processing

Medical image segmentation is a strategy for extricating the ideal parts and highlights from the info medical image information. The presentation of classification stage depends on initial stages like preprocessing and segmentation. The traditional Fuzzy c-means (FCM) clustering algorithms have been generally utilized for grayscale and color image segmentation. In this work, we propose a super-pixel based FCM clustering algorithm that is altogether more hearty than best in clustering algorithms for image segmentation. We initially acquire a preprocessing stage by super-pixel image with exact contour for background separation. As opposed to customary neighboring window of fixed size and shape, the super-pixel image gives better adaptive and irregular local spatial neigh-borhoods that are helpful for improving Interstitial lung disease (ILD) image segmentation. Also after that the results are compared with preprocessing performed by adaptive median filtering to stay away from the noise effect on ILD images followed by Contrast-limited adaptive histogram equalization (CLAHE) enhancement to improve the image quality and then segmented by FCM. The outcomes are obtained for various number of clusters segmented with FCM with super-pixel approach and result are improve as contrast to conventional FCM and Otsu method on ILD images.