Abstract
Image segmentation is an important problem in image processing and object recognition, and is one well-known bottleneck for further applications. Fuzzy C-means, as one typical clustering algorithm in pattern recognition, has been improved for image segmentation in many aspects. Aiming at the distance form in FCM, this paper proposes to incorporate FCM with kernel functions, which will make it insensitive to noise and other artifacts. Experiments show that the proposed algorithm can retrieve better results than other improved algorithms.