Robust interpolation of DT-MRI data using Tensor Splines
Barmpoutis A, Shepherd T, Vemuri B, Forder J
University of Florida
Interpolation is a fundamental problem in the field of medical imaging. To correctly perform interpolation of diffusion tensors one needs to first identify the space in which these diffusion tensors live and then define an appropriate metric in this space for interpolation and computation of lengths of paths, averages etc. The space of diffusion tensors is known to be a curved space of negative curvature. We make use of this to define an appropriate metric to perform interpolation and other tasks. Interpolation is essential in many applications such as image registration, atlas construction etc.