UBC Theses and Dissertations
Radon Transform in three dimensional image reconstruction from projections Stazyk, Michael Walter
This thesis presents an algorithm for image reconstruction from projections intended for use in a new class of volume imaging PET scanners. The algorithm is based on the inversion of the three dimensional Radon Transform as it applies to the truncated cylindrical detector geometry and is derived from the X-ray Transform inversion given by the Orlov recovery operator. The algorithm is tested using Monte Carlo simulations of several phantom geometries and employs a single iterative step to include all detected events in the reconstruction. The reconstructed images are good representations of the original objects, however the iterative step is a source of some significant artefacts in the images. Also discussed is the extension of the Radon Transform technique to a non-iterative method for three dimensional image reconstruction using all detected events.
Item Citations and Data