Multi-resolution 3D Nonrigid Registration via Optimal Mass Transport on the GPU

In this paper we present computationally efficient imple- mentation of the minimizing flow approach for optimal mass transport (OMT) with applications to non-rigid 3D image registration. Our imple- mentation solves the OMT problem via multi-resolution, multigrid, and parallel methodologies on a consumer graphics processing unit (GPU). Although computing the optimal map has shown to be computationally expensive in the past, we show that our approach is almost two orders magnitude faster than previous work and is capable of finding transport maps with optimality measures (mean curl) previously unattainable by other works (which directly influences the accuracy of registration). We give results where the algorithm was used to compute non-rigid regis- trations of 3D synthetic data as well as intra-patient pre-operative and post-operative 3D brain MRI datasets.