Accepted for Presentation (Paper 2050), 1996 International Conference on Parallel Processing, Bloomingdale, Illinois, 12-16 August 1996.
Three dimensional computed tomography is a computationally intensive procedure, requiring large amounts of RAM and processing power. Parallel methods for two dimensional computed tomography (CT) have been studied, but not extended or applied to 3D cone beam tomography. Parallel methods for iterative CT algorithms have been studied, but only on small problems. After a review of these and related topics, a discussion on how the 3D cone beam tomography problem can be performed in parallel is given. Results of an MPI implementation of the voxel driven backprojection on an SMP and cluster of workstation show 80% efficiency.