Fatemeh Saberi - Applied and science university center khane kargar, gorgan , golestan ,iran

Kernel fuzzy c-means clustering algorithmuse the kernel distance and spatial error onits membership functions to realize objectivefunction in Fuzzy c-means algorithm inimage segmentation. This algorithm isimplemented through Euclidian distanceconversion to kernel distance characteristicson property space. Membership criterionand the sequences of cluster centroidsequations are achieved by minimizing theefficient objective functions; the centralinitializing algorithm is presented so that todecrease computational and time complexity. The proposed algorithm implement onblack & white, colored and noisy medicalimages in MATLAB software with differentclusters which is more precise and strongerthan Fuzzy c-means algorithm with highersimilarity.

Image segmentation, kernel distance , error spatial,objective function,fuzzy c-means