A modified-FCM segmentation algorithm for brain MR images
Ruoyu Du, Hyo Jong Lee · 2009
Medical image segmentation is a complex and challenging task due to the intrinsic nature of the images. Segmentation of magnetic resonance imaging (MRI) is widely used in medical area. One of common clustering algorithms is Fuzzy c-means (FCM) for segmentation of MR images. However MR images normally significant noise by the impact of principles of imaging, equipment and environment, which can lead to serious inaccuracies with segmentation. Based on the FCM clustering algorithm, an improved segmentation technique is proposed in this paper. The Sigma filter principle has been applied to change the neighbor pixels of targets. The efficacy of the proposed algorithm is demonstrated by comparison with FCM algorithm in visual evaluation and quantitative evaluation.