Separation of Gray and White Matter from MR Images Using the Polya Urn Model and the EM Algorithm
Faguo Yang, Tianzi Jiang, Ieee Member · 2014
In this paper, a novel algorithm for the separation of gray and white matter from single sequence magnetic resonance images is proposed. In our approach, the Polya urn model is used to model the smoothness and contiguous nature of the tissue regions. The order of the neighborhood system used in the Polya urn model is adaptive with the contents of the image. Moreover, an adaptive window is used to resist the intensity inhomogeneity of the magnetic resonance images, when estimating the parameters of each cluster using the expectation-maximization (EM) method. Experimental results demonstrate that our approach can extract gray and white matter from magnetic resonance images quickly and exactly. 1.