Research Of Image Segmentation Technology Based On Fuzzy C-Means Clustering

With the popularization of computer, the development of Internet and Multimedia Technology, people utilize computer to get and manage the visual image information. In order to distinguish and analyze the interested region, Image Segmentation has emerged to be one of the hot research areas in image domain.In this dissertation, lots of exploratory research work has been done around some deficiencies of Fuzzy C-Means Clustering (FCM) in Image Segmentation. The main contributions of this dissertation are summarized as follows:⑴An adaptive FCM image segmentation algorithm based on the feature divergence is proposed aiming at the deficiencies of the traditional FCM, such as Euclidean distance, stochastically setting the initial number of clusters, and so on. This algorithm can accomplish image segmentation by importing the feature divergence, extracting the image feature according to Laws texture measure, incorporating the cluster validity exponent to ascertain the initial number of clusters adaptively. Experimental results show that the proposed method is simple and efficient (especially for texture images), its global performance is superior to the existing FCM image segmentation schemes.(2) An algorithm for color image segmentation using ReliefF and FCM is proposed taking the different effects of each pixel in an image into consideration. Firstly, hue,intensity,saturation (HIS) space is selected which approaches the human vision system. Then, the ReliefF are adopted for calculating the weight of each pixel, restricting the membership sequentially, and the intension of punish effect is controlled by the weighted factors of punishment, thereby the robustness research on FCM is accomplished. Experimental results show that the proposed scheme is more efficient in segmenting color images with noise.