OR-EM Method for Image Reconstruction in Cone-Beam CT

Abstract

Traditional CT reconstructions are limited by many kinds of artifacts, and they give dissatisfactory image. To reduce image noise and artifacts, we propose a statistical iterative method for image reconstruction in cone-beam CT. First the theory of maximum likelihood estimation is extended from emission CT to X-ray scan. Then an expectation-maximization (EM) formula is deduced for direct reconstruction of cone-beam CT. EM algorithm is an iterative method that can produce good quality reconstruction. But compared with fast and robust FDK algorithm, EM algorithm is computer intensive and convergence slow. In order to accelerate the convergence speed of EM algorithm, over relaxation (OR) is applied in this work. OR-EM algorithm is a feasible iterative method. It can provide good quality reconstructions after only a few iterations. Experiments results with computer simulated data and real CT data show OR-EM algorithm is effective. Its reconstructed images are accurate and clear.