A Novel Fuzzy C-Means for Image Segmentation
In this paper, we present a novel algorithm for fuzzy segmentation of infrared image data using fuzzy clustering. A conventional FCM assigns the data into group, where the data is nearest to the center of group. Although FCM is populated in image segmentation, it still has the following disadvantages: (1) a conventional FCM algorithm does not consider spatial information for clustering. (2) The algorithm is sensitive to noise. In this paper we present a fuzzy-means algorithm that incorporates spatial information and the prior probability of a pixel neighborhood into the membership function for clustering. The modified FCM has a great improvement for noisy image and infrared image segmentation.
KeywordsFuzzy k-means Spatial information Clustering Segmentation
The authors would like to thank the anonymous reviewers for their detailed review and constructive comments. This work is supported by the National Natural Science Foundation of China (Grant no. 60736046).
- 1.Dunn JC, Cybernet J (1973) A fuzzy relative of the ISODTA TA process and its use in detecting compact well separated clusters [J] 3(3):32–57Google Scholar
- 2.Pattern recognition with fuzzy objective function algorithms [M], Plenum Press, New York (1981) pp 121–122Google Scholar
- 3.Chuang K-S, Tzeng H-L (2006) Fuzzy c-means clustering with spatial information for image segmentation [J]. Comput Med Imaging Graph 30:9–15Google Scholar
- 6.Saha S, Bandyyopadhyay S (2007) MRI brain image segmentation by fuzzy symmetry based genetic clustering technique [C]. IEEE Congr Evol Comput 4417–4424Google Scholar
- 9.Bo Yuan, George J. Klir (1995) Evolutionary fuzzy c-means clustering algorithm [C].IEEE 2221–2226Google Scholar