These are the results of my involvement in a research project with Dr. Thomas Humphries at UW Bothell. We were researching improvements to the iterative algorithm SART for reconstructing low-dose CT images. Our experiments revolved around utilizing a non-differentiable function (BM3D) to generate perturbations for the superiorization of SART. More thorough background can be found in the accompanying paper, BM3D as a Superiorization Function Within SART for Denoising of Computed Tomography Images (embedded below). The .tex file is available for inspection here. Also included in this branch is a .yml file which should be capable of setting-up a conda environment suitable for running the software. However, some of the requirements in the .yml file may be hardware-dependent. For comparison with preexisting work, please see the original branch of this project, from Dr. Humphries.
