PARTIAL-VOLUME BAYESIAN CLASSIFICATION OF MATERIAL MIXTURES IN MR VOLUME DATA USING VOXEL HISTOGRAMS

We present a new algorithm for identifying the distribution of differe nt material types in volumetric datasets such as those produced with m agnetic resonance imaging (MRI) or computed tomography (CT). Because w e allow for mixtures of materials and treat voxels as regions, our tec hnique reduces errors that other classification techniques can create along boundaries between materials and is particularly useful for crea ting accurate geometric models and renderings from volume data, It als o bras the potential to make volume measurements more accurately and c lassifies noisy. low-resolution data well. There are two unusual aspec ts to our approach. First, we assume that, due to partial-volume effec ts, off blurring, voxels can contain more than one material, e.g., bot h muscle and fat; we compute the relative proportion of each material in the voxels. Second, we incorporate information from neighboring vox els into the classification process by reconstructing a continuous fun ction, rho(x), from the samples and then looking at the distribution o f values that rho(x) takes on within the region of a voxel. This distr ibution of values is represented by a histogram taken over the region of the voxel; the mixture of materials that those values measure is id entified with the voxel using a probabilistic Bayesian approach that m atches the histogram by finding the mixture of materials within each v oxel most likely to have created the histogram. The size of regions th at we classify is chosen to match the spacing of the samples because t he spacing is intrinsically related to the minimum feature size that t he reconstructed continuous function can represent.