An automated MRI brain tissue segmentation approach
Zaineb Ben Messaoud, Siwar Chnitti, Ines Njeh · 2016
Accurate magnetic resonance brain tissue segmentation is of much importance in medical imaging. Hence segmentation methods are in research focus and various methods are presented in the literature. In this paper, a multi region graph cut image segmentation in a kernel-induced space is used for brain-tissue-segmentation framework. The RBF kernel function transforms implicitly image data so that the piecewise constant model of the graph cut formulation becomes applicable. A preprocessing approach was proposed for automatic whole-brain extraction from MR images. Evaluation over the publicly training data set from the framework for MR Brain Image Segmentation (MRBrainS13) proved that our method achieves a competitive performance for brain tissue segmentation, among existing methods.