Color Image Segmentation Using Kernalized Fuzzy C-means Clustering

The fuzzy c-means (FCM) algorithm is a very popular algorithm in the field of image segmentation because of its simplicity and less sensitivity to noise and it is widely used in the field of engineering disciplines. The FCM membership function can handle the overlapped clusters efficiently with predefined number of clusters, but this algorithm are unable to cluster non-linearly separable data as well as choosing of initial cluster centre is difficult task which results in poor image segmentation. To overcome this drawback, we proposed Kernalized Fuzzy C-means (KFCM) clustering. In that kernel space clustering is used for clustering of nonlinear image, which have kernel functions which transform data in image plane into higher dimension feature space and these kernel functions are used to find non-Euclidean distance between feature point without defining transfer function, and then perform FCM in feature space. Here we use two different kernel functions for image segmentation and compare their outputs.