Hidden-Data Spaces for Maximum-Likelihood PET Reconstruction

The author shows that expectation-maximization (EM) algorithms based on smaller complete data spaces will typically converge faster. As an example, he compares the two maximum-likelihood (ML) image reconstruction algorithms of D. G. Politte and D. L. Snyder (1991) which are based on measurement models that account for attenuation and accidental coincidences in positron-emission tomography (PET).Peer Reviewedhttp://deepblue.lib.umich.edu/bitstream/2027.42/86022/1/Fessler120.pd