An efficient automatic framework for segmentation of MRI brain image

A full automatic framework for segmentation of brain image is proposed in this paper. The method is able to segment MRI images, corrects for MRI signal inhomogeneities, and incorporates contextual information by means of Markov random filed. The framework consists of three-step segmentation procedures. First, non-brain structures removal by level set method. Then, the non-uniformity correction method is based on computing estimates of tissue intensity variation. Finally, it uses a statistical model based on Markov random filed (MRF) for MRI brain image segmentation. The brain tissue can be classified into cerebrospinal fluid (CSF), white matter (WM) and gray matter (GM). The efficacy of the proposed method is demonstrated by extensive segmentation experiments using both simulated and real MR images.