Fuzzy Image Segmentation using Suppressed Fuzzy C-Means Clustering (SFCM)

Ali, Ameer; Karmakar, Gour C. and Dooley, Laurence S. (2004). Fuzzy Image Segmentation using Suppressed Fuzzy C-Means Clustering (SFCM). In: International Conference on Computer and Information Technology, 26-28 Dec 2004, Dhaka, Bangladesh.

Abstract

Clustering algorithms are highly dependent on the features used and the type of the objects in a particular image. By considering object similar surface variations (SSV) as well as the arbitrariness of the fuzzy c-means (FCM) algorithm for pixel location, a fuzzy image segmentation considering object surface similarity (FSOS) algorithm was developed, but it was unable to segment objects having SSV satisfactorily. To improve the effectiveness of FSOS in segmenting objects with SSV, this paper introduces a new fuzzy image segmentation using suppressed fuzzy c-means clustering (FSSC) algorithm, which directly considers object SSV and incorporates the use of suppressed-FCM (SFCM) using pixel location. The algorithm also perceptually selects the threshold within the range of human visual perception. Both qualitative and quantitative results confirm the improved segmentation performance of FSSC compared with other algorithms including FSOS, FCM, possibilistic c-means (PCM) and SFCM for many different images.