I. Introduction
Diffusion Tensor Imaging (DTI) measures local probability distribution of the self-motion of water molecules. Restricted motion of extra-and intra-cellular fluid within brain white matter fibers correlates local diffusion patterns, in its simplest form represented by local tensors, with the orientation and density of white matter fibers. The potential of DTI to present white matter integrity, disruption and pathology long before visible changes occur in structural imaging makes it the preferred modality to study white matter diseases. The two tensor measures most commonly used in clinical analysis are the “apparent diffusion coefficient” (ADC, trace of tensor) and the “fractional anisotropy” (FA, normalized difference from spherical) Previous research has been mostly focused on the robust calculation of the tensor field, regularization [1], scientific visualization [2] and fiber tracking [3], [4], [5], [6], [7], [8]. Atlas building and comparison of subject groups have been studied by non-linear registration of DTI images [9] and by combining tractography and spatial normalization [10].