MRI segmentation using Fuzzy C-means and radial basis function neural networks

Image segmentation is one of the major preprocessing steps of magnetic resonance imaging (MRI) analysis in many medical and research applications. Accurate differentiation between three major soft tissues of the brain - grey matter, white matter and cerebrospinal fluid - is a key step in structural and functional brain analysis, visualization of the brain's anatomical structures and measurement, diagnosis of neurodegenerative disorders and image-guided interventions as well as surgical planning. We propose a new methodological approach in segmentation of MRI images of the brain structure. Although various methods for MRI segmentation have been proposed, improvement of soft, automatic and precise MRI segmentation methods are worth a try. The proposed method has almost the same results as those from recent efforts in this field. However, it performs better in the presence of noise and RF-filed inhomogeneity.