Multispectral MRI image segmentation using Markov random field model

Magnetic resonance imaging (MRI) is used to capture images in different modalities such as T1-weighted, T2-weighted, and PD-weighted. This paper proposes a new method for the fusion of different channels in MRI image segmentation. In the reported work, a new feature vector for multispectral MRI brain segmentation is proposed. Fuzzy C-means clustering method is applied on the three different extracted feature vectors, and results are reported. Experimental results show that the proposed feature vector presents good noise immunity. Paper reports a new segmentation method based on Markov random field and the proposed feature vector to combine spatial and spectral information for MRI image segmentation. The proposed method was applied on the BrainWeb MRI image dataset with added noise, and the segmentation results are reported and compared with some known reported works.