Infrared image segmentation based on fast fuzzy C-means clustering

The fuzzy C-means (FCM) algorithm has many disadvantages such as number of clusters must be determined before FCM clustering is implemented and the algorithm needs an amount of calculation. In order to solve these problems, a novel method of fast FCM clustering is proposed. Seed pixels can be obtained by neighborhood searching of edge information firstly; Number of clusters and the value of cluster centers can be achieved by region growing method. Image is separated into cluster regions and undetermined cluster regions. The value of cluster centers and FCM are adopted to determine the undetermined cluster regions. Experiences show that the new method greatly improved the efficiency of image segmentation. Since the relationship of neighbored pixels are taken into account, the results of ima