Multi-channel registration of diffusion tensor images using directional information

The problem of registering diffusion tensor (DT) images is considered. We describe a novel intensity based registration method capable of performing affine and nonlinear registration of multi-channel images such as DT images. We use this method to register 3 dimensional DT images of the human brain based on several channel configurations derived from the DT model. Specifically, we compare the use of channel configurations that include rotationally invariant scalar quantities derived from the DT model against channel configurations that include directional information, such as the elements of the diffusion tensor. Experiments performed with real and simulated data show that the use of the directional information present in the diffusion tensor elements can, in some instances, significantly improve the accuracy of the registration results when compared to methods that use rotationally invariant scalar information.