Effective FCM noise clustering algorithms in medical images S.R. Kannan a,n , R. Devi a , S. Ramathilagam b , K. Takezawa c a Department of Mathematics, Pondicherry Central University, India b Department of Mathematics, Periyar Government College, Tamil Nadu, India c NARC, Tsukuba, Japan article info Article history: Received 19 May 2011 Accepted 21 October 2012 Keywords: Noise clustering Entropy FCM Medical images Segmentation abstract The main motivation of this paper is to introduce a class of robust non-Euclidean distance measures for the original data space to derive new objective function and thus clustering the non-Euclidean structures in data to enhance the robustness of the original clustering algorithms to reduce noise and outliers. The new objective functions of proposed algorithms are realized by incorporating the noise clustering concept into the entropy based fuzzy C-means algorithm with suitable noise distance which is employed to take the information about noisy data in the clustering process. This paper presents initial cluster prototypes using prototype initialization method, so that this work tries to obtain the final result with less number of iterations. To evaluate the performance of the proposed methods in reducing the noise level, experimental work has been carried out with a synthetic image which is corrupted by Gaussian noise. The superiority of the proposed methods has been examined through the experimental study on medical images. The experimental results show that the proposed algorithms perform significantly better than the standard existing algorithms. The accurate classifica- tion percentage of the proposed fuzzy C-means segmentation method is obtained using silhouette validity index. & 2012 Elsevier Ltd. All rights reserved. 1. Introduction Diagnostic imaging is a precious tool in medicine nowadays. There are different medical image acquisition techniques such as magnetic resonance imaging, ultrasound, X-ray computer tomo- graphy, single photon emission tomography, positron emission tomography etc. MRI [19,22] is a superior, commonly used medical imaging technique. It provides detailed images of living tissues and has advantages over other imaging techniques for both brain and breast studies [12]. The registration of magnetic resonance images of the brain and breast has recently become an extensively used research tool [1,4] with the potential to enter usual medical use. In the analysis of medical images for diagnos- ing the diseases, segmentation is often essential as a preliminary stage. Successful numerical outcome in segmented medical image can assist physicians to study and detect the structure and function of the body in both health and disease. Accurate segmentation is necessary for detecting the diseased spot in medical images. Human error occurred while segmenting the medical images manually due to the intrinsic nature of the images. Also manual segmentation is a challenging and time consuming task. Therefore, computer aided segmentation is very significant to find out effective results in medical images. The brain and breast organs in human body have a convoluted structure and its computer aided segmentation is very essential for detecting normal and abnormal areas, in order to follow appropriate treat- ment. But, the medical images always comprise substantial ambi- guity, unknown noise, partial volume effect and intensity inhomogeneity; these imaging artifacts degrade the segmentation process. The noises and other imaging artifacts cause serious misclassification and overlapping tissues when the computer aided analysis in medical image segmentation [6,11,26]. Selection of a suitable approach to a segmentation problem can be a complicated dilemma. Clustering is a popular unsupervised classification method and has found many applications in pattern classification and image segmentation. Clustering algorithm attempts to classify a voxel to a tissue class by using the notion of similarity to the class. Since there is no information given about the underlying data structure or the number of clusters, there is no single solution to clustering; neither is there a single similarity measure to differentiate all clusters. Therefore for this reason there is no theory which illustrates cluster- ing uniquely. This paper has used the paradigm of fuzzy clustering which is based on the elements of fuzzy set theory for segmenting medical images. Fuzzy based clustering methods [2,3,9,10,21,30,32] have attracted more attention for image segmentation techniques, because they gathered more information from the image [26]. Because of Euclidian distance based objective function of standard FCM, it works well in clustering the noise free data and it fails to Contents lists available at SciVerse ScienceDirect journal homepage: www.elsevier.com/locate/cbm Computers in Biology and Medicine 0010-4825/$ - see front matter & 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.compbiomed.2012.10.002 n Corresponding author. Tel.: þ919865707773. E-mail address: srkannan_gri@yahoo.co.in (S.R. Kannan). Computers in Biology and Medicine 43 (2013) 73–83