Segmentation methods of Magnetic Resonance Images based on Markov Random Fields

摘要

The aim of this paper is to compare methods of image segmentation based on Markov Random Fields to biomedical images coming from Magnetic Resonance Imaging (MRI). For the optimization of energy function four algorithms were used: Metropolis, Gibbs Sampler, Iterated Conditional Modes (ICM), and Modified Metropolis Dynamics (MMD). As a result of the segmentation the activity parts of the brain from fMRI are shown. Moreover, some pathologies like brain cancers are labeled in MRI images.