Accelerated image reconstruction using variable weight ordered subsets of projection data

Abstract

This paper study OSEM image reconstruction algorithm on different subset of a series of numerical simulated data, through the mean square error of image reconstruction on different subset in the same generation of iteration, analysis effect of reconstructive image and convergence rate. Ordered subsets EM (OS-EM) provides a restoration imposing a natural positivity condition and with close links to the EM algorithm. OS-EM is applicable in both single photon (SPECT) and positron emission tomography (PET). In simulation studies in SPECT; the OS-EM algorithm provides an order-of-magnitude acceleration over EM, with restoration quality maintained. We define variable weight ordered subset processing for standard algorithms for image restoration from projections. Variable weight ordered subsets methods group projection data into an ordered sequence of subsets (or blocks). An iteration of ordered subsets EM is defined as a single pass through all the subsets, in each subset using the current estimate to initialize application of EM with that data subset. Reconstruct image with variable weight subset (decreasing subset) accelerated image reconstruct. Good image with smaller iterative and thus shorter time can be obtained by properly sequence subset.