Segmentation of Brain Magnetic Resonance Images

This chapter presents the application of different rough-fuzzy clustering algorithms for segmentation of brain magnetic resonance (MR) images. One of the important issues of the partitive-clustering-algorithm-based brain MR image segmentation method is the selection of initial prototypes of different classes or categories. The concept of discriminant analysis, based on the maximization of class separability, is used to circumvent the initialization and local minima problems of the partitive clustering algorithms. The chapter first deals with the pixel classification problem, and then gives an overview of the feature extraction techniques employed in segmentation of brain MR images, along with the initialization method of c-means algorithm based on the maximization of class separability. It presents implementation details, experimental results, and a comparison among different c-means algorithms. The algorithms compared are hard c-means (HCM), fuzzy c-means (FCM), possibilistic c-means (PCM), FPCM, rough c-means (RCM), and rough-fuzzy c-means (RFCM).

fuzzy set theory; image classification; image segmentation; magnetic resonance imaging; pattern clustering; rough set theory