Fast ordered subset reconstruction for X-ray CT

Statistical iterative methods for image reconstruction like Maximum Likelihood Expectation Maximization (ML-EM) are more robust and flexible than analytical inversion methods and allow for accurately modeling the counting statistics and the photon transport during acquisition. Up to recently, statistical reconstruction algorithms were prohibitively slow when applied to clinical X-ray CT due to the large data sets and the high number of iterations required for reconstructing high resolution images. Recently, powerful acceleration methods for statistical reconstruction based on using ordered subsets (OS) of projection data have been proposed. In this paper we study images generated by an OS accelerated algorithm, the OS convex algorithm (OSC), for data sets with sizes, noise levels and spatial resolution representative for X-ray CT imaging. In the case of only a few projections per subset, areas with decreased intensity appear in the OSC reconstructed images, which can be adequately corrected for by running the final iteration with a reduced number of subsets. Even then OSC reaches an equal resolution more than two orders of magnitude faster than the standard convex algorithm