Research On Image Segmentation Based On Fuzzy Clustering Algorithm

Image segmentation technology has gained a lot of people's attention.Researchers have made countless methods of image segmentation for many years.Image segmentation technology has been involved in many fields,and becomes the basis of many technologies gradually.Image segmentation technology still has a very important position in current era.How to improve the image segmentation results still have very important research significance.Among many methods of image segmentation,there is a kind of image segmentation method based on fuzzy clustering which has become the most widely used segmentation technology and hot research topic because of its ability to deal with the inherent ambiguity.Many of the things themselves are ambiguous,the traditional clustering method is hard division and can not handle such things ideally.Fuzzy clustering analysis is a typical soft partitioning method,which can deal with the inherent ambiguity of things,so it is more suitable for practical things than traditional clustering analysis.Among the algorithms of fuzzy clustering analysis,the most widely used algorithm is FCM algorithm.Although the FCM algorithm is the most widely used,it still has some problems,so the researchers have improved it and putting forward a lot of improved algorithms.In recent years a typical improvement algorithm is called KWFLICM algorithm.KWFLICM algorithm is an important algorithm based on FLICM,and it was proposed which combined with the idea of kernel function proposed for further improvement.This article has made improvements on the deficiency of KWFLICM algorithm.First,this paper supplements some contents of which the KWFLICM algorithm is not clearly described.Then,designed a new fuzzy factor,and the algorithm segmentation effect and anti-noise result have been further improved.In the part of experiment,a lot of images are applied to the method,including synthesis images,medical images and natural images.The results of the experiments show that the improvements in this paper are valid.