Color image segmentation based on improved FCM algorithm incorporating spatial information

Based on spatially weighted FCM considering the spatial information of pixels, we propose a novel algorithm which can get faster and more accuracy color image segmentation results. We convert RGB color image features into HIS space, and use improved FCM clustering algorithm to get membership function of I component of pixels. Finally, we form a new feature with the membership function values and H component of pixels, and use Kmeans for final clustering. The results of experiments show the effectiveness of the proposed algorithm.