Image segmentation based on fast kernelized fuzzy clustering analysis

Based on kernelized fuzzy clustering analysis, this paper presents a fast image segmentation algorithm using a speeding-up scheme called reduced set representation. The proposed clustering algorithm has lower computational complexity and could be regarded as the generalized version of the traditional KFCM-I and KFCM-II algorithms. Moreover, an image intensity correction is employed during image segmentation process. With another speeding-up scheme called pre-classification, the proposed intensity correction could further acclerate image segmentation. Experiments of MRI image segmentation have shown the effectiveness of the proposed algorithm, which outperforms in its rivals.