Reorientation strategies for High Order Tensors

In Diffusion Magnetic Resonance Imaging, the diffusion process is commonly modeled by a symmetric 2nd order tensor. Since 2nd order tensors are not able to properly model crossing fibers, High Order Tensors (HOTs) can be considered in particular to model the Orientation Distribution Function. Applying a spatial transformation to such tensor fields requires an appropriate reorientation strategy to preserve the relevancy of the orientational information related to fiber organization. We propose a general framework for HOT reorientation, relying on the decomposition of an HOT as a function of several 2nd order tensors. Contrary to existing methods, this approach enables to reorient fibers from a crossing independently. Experiments on simulated data highlight the benefit of the proposed method.