Enhanced Spatial Priors for Segmentation of Magnetic Resonance Imagery

A framework for probabilistic segmentation of Magnetic Resonance (MR) images is proposed which utilizes three types of models: intensity models to capture the graylevel appearance of a structure, relative-spatial models which describe the spatial relationships between structures in a subject-specific reference frame, and shape models to describe the shape of structures in a subjectindependent reference frame. A vast literature exists on intensity as well as shape models, and the main contribution of this work is the development of two relative-spatial models. A discrete vector valued Markov Random Field (MRF) model, whose parameters are derived from training data, is used to capture the piecewise homogeneity (white matter is likely to occur next to white matter) and the neighborwise compatibility (white matter is unlikely to occur next to skin) of different tissues. The continuous Mean Field solution to the MRF is recovered using ExpectationMaximization algorithm, and is a probabilistic segmentation of the image. The second model is a conditional spatial distribution, also created using training data, on specific structures (such as cartilage, or brain tissue) conditioned on the geometry of other structures (such as trabecular bone, or ventricles). The motivation is to bootstrap the segmentation process using spatial relationships to structures that image well using MR and hence are segmented easily. Results are presented for the segmentation of white matter, gray matter, fluid, and fat in Gradient Echo MR images of the brain, and for the segmentation of trabecular bone in T2-weighted MR images of the knee.