Improving Reconstructed Image Quality in a Limited-Angle Positron Emission Tomography System

Limited-angle positron emission tomography systems can be placed next to the anatomical area of interest, improving their spatial resolution and diagnostic sensitivity. However, the incomplete angular sampling makes the reconstruction process susceptible to image artifacts and anisotropic spatial resolution. In this work, I show that combining total variation (TV) minimization with the maximum likelihood expectation maximization (MLEM) algorithm in the image reconstruction process can lead to greatly improved image quality for datasets with limited angular sampling range.