Bayesian Reconstruction Algorithm for PET Using New Markov Quadratic Hybrid Multi-Order Priors

Bayesian reconstruction, or maximum a posteriori (MAP) method, has proved its superiority over others in positron emission tomography (PET) image reconstruction. In this article, a novel type of concave MRF (Markov random fields) hybrid convex priors for Bayesian reconstruction, which combine quadratic smoothness priors of different orders, are proposed. The design of the new priors is based on the intrinsic properties of the smoothness priors of different orders and aims to make an adaptive use of the smoothness priors. Effective parameter estimations are given Simulation experiments of their application in PET reconstruction are illustrated. Results and comparisons proved the new hybrid priors´ good performance in lowering noise effect and preserving edges.