An Improved FCM Algorithm And Its Application In Image Segmentation

FCM algorithm is also called fuzzy C-means(Fuzzy C-means)algorithm,which is a fuzzy clustering algorithm based on objective function.There are two problems in the FCM algorithm,one is the clustering validity problem,the other is the clustering center initialization problem.On the one hand,in the FCM algorithm,because the number of clusters is unknown,it is generally necessary to give in advance that the quality of clustering is closely related to the number of clusters.On the other hand,when clustering with FCM algorithm,it is considered that the effect of each sample point on the clustering result is the same,and the contribution of each sample point to clustering can not be given reasonably.Therefore,based on the traditional FCM algorithm,this paper proposes an improved weighted FCM validity function,and gives the iterative formula of clus-tering center and weighting parameters by Lagrangian multiplication.Secondly,for clustering validity In this paper,we propose a method to determine the opti-mal clustering number C method by using the effective peak number of the image histogram.Finally,the FCM algorithm can not reasonably give the problem of the size of the contribution of each sample point to the cluster,This paper establishes the contribution of each sample to the clustering results by using the frequency of each sample in each class for each sample in the class.It is found that the improved weighted FCM algorithm is very useful for determining the clustering center and weighting parameters,and it is found that the weighted FCM algo-rithm can effectively reduce the iterative speed and iteration time and reduce the noise.Gray image segmentation is very good,and to ensure the correctness of the cluster.