NRFCM: A New Robust Fuzzy Clustering Algorithm for Image Segmentation
Artificial Intelligence and Computational Intelligence, 2009. AICI '09. International Conference（2009）
摘要
The fuzzy c-means algorithm (FCM) has been proven effectively for image segmentation. RFCM is an improvement algorithm of FCM. However, RFCM still has the following disadvantages: (1) RFCM cannot effectively avoid the impact of noises; (2) In RFCM, the noise is regarded as the normal sample and RFCM does not smooth the noise point without considering the relationship between the noise and its neighborhood. In this paper, by incorporating local spatial and gray information, a new robust fuzzy clustering algorithm for image segmentation (NRFCM) is proposed. The major characteristics of NRFCM are as follows: (1) We can effectively reduce the negative influence of the noise on the clustering results by using a new factor, which is a penalty on the distance. (2) The block noises have been avoided by bringing in the cluster weight, which is represented the priori probability of clusters. Experiments show that NRFCM is more suitable for image segmentation by comparing with RFCM, FASTFCM and FCMS_1.
更多查看译文
关键词
nrfcm,clustering result,pattern clustering,cluster weight,image,rfcm,image segmentation,improvement algorithm,fcm,fuzzy c-means clustering(fcm),noise information,robust fuzzy clustering algorithm,block noise,noise point,gray information,fuzzy c-means algorithm,following disadvantage,new robust fuzzy clustering,new factor,pixel,data mining,fuzzy clustering,noise,prototypes,robustness,clustering algorithms
AI 理解论文
溯源树
样例
生成溯源树，研究论文发展脉络
Chat Paper
正在生成论文摘要