Cerebral magnetic resonance image segmentation using fuzzy Markov Random Fields

In this paper, we present a fuzzy Markovian method for brain tissue segmentation from magnetic resonance images. Generally, there are three principal brain tissues in a brain dataset: gray matter, white matter and cerebrospinal fluid. However, due to the limited resolution of the acquisition system, many voxels may be composed of multiple tissue types (partial volume effects). The proposed method aims to calculate the fuzzy membership of each voxel to indicate the partial volume degree using a fuzzy Markovian segmentation. Since our method is unsupervised, it first estimates the fuzzy Markovian Random Field model parameters using a stochastic gradient algorithm. The efficiency of the proposed method is quantified on a digital phantom using an absolute average error, and qualitatively tested on real MRI brain data. The results of the fuzzy segmentation can be applied to recognition of internal structures or segmentation of tumor.