Magnetic resonance image tissue classification using a partial volume model

We describe a sequence of low-level operations to isolate and classify brai n tissue within T1-weighted magnetic resonance images (MRI). Our method fir st removes nonbrain tissue using a combination of anisotropic diffusion fil tering, edge detection, and mathematical morphology. We compensate for imag e nonuniformities due to magnetic held inhomogeneities by fitting a tricubi c B-spline gain field to local estimates of the image nonuniformity spaced throughout the MRI volume. The local estimates are computed by fitting a pa rtial volume tissue measurement model to histograms of neighborhoods about each estimate point. The measurement model uses mean tissue intensity and n oise variance values computed from the global image and a multiplicative bi as parameter that is estimated for each region during the histogram fit. Vo xels in the intensity-normalized image are then classified into six tissue types using a maximum a posteriori classifier. This classifier combines the partial volume tissue measurement model with a Gibbs prior that models the spatial properties of the brain. We validate each stage of our algorithm o n real and phantom data. Using data from the 20 normal MRI brain data sets of the Internet Brain Segmentation Repository, our method achieved average kappa indices of kappa = 0.746 +/- 0.114 for gray matter (GM) and kappa = 0 .798 +/- 0.089 for white matter (WM) compared to expert labeled data. Our m ethod achieved average kappa indices kappa = 0.893 +/- 0.041 for GM and kap pa = 0.928 +/- 0.039 for WM compared to the ground truth labeling on 12 vol umes from the Montreal Neurological Institute's BrainWeb phantom. (C) 2001 Academic Press.