Research For Image Segmentation Based On Fuzzy Clustering Analysis

Image segmentation is just to segment an image into different sub-images with different characters and get some interested objects. It is a key step from image process to image analysis, and is applied in a lot of fields such as computer vision, image coding, pattern recognition, medical image processing and so on.Images themselves are very uncertain and inaccurate. In recent years, some experts are making efforts to apply the fuzzy clustering method to image segmentation, and it is more effective than the traditional image processing method. The paper first introduce the basic theory and methods of fuzzy math, leads to hard C-means algorithm and fuzzy C-means algorithm from data sets to the delineation, and then improve the inadequacy of the Fuzzy C-means clustering algorithm application in the image segmentation. And then demonstrates the definition of image segmentation and existing several of image segmentation algorithm, apply the improved fuzzy clustering algorithm to image segmentation.Improve it in two areas for the shortcomings of FCM classical algorithm:First, through increase the weight index in the objective function of fuzzy object to reduce the effect of sample vector to the cluster result; the second is by improving fuzzy membership to overcome the relationships of the image pixels around the cluster centre.The paper apply the improved FCM algorithm to image segmentation. The paper firstly discussed the image segmentation of FCM algorithm based on the neighborhood pixels, and then from the fuzzy membership bound point of view, proposes a new cluster objective function and an improved FCM image segmentation algorithm. Experiments show that the new algorithm has a good image segmentation effect.