Locations of discontinuities as priors for improved Bayesian reconstructions from projections (emission tomography)

Abstract

A Bayesian image reconstruction method from projection data using explicit discontinuity locations as priors is proposed. The algorithm seeks smooth solutions with occasional discontinuities with a special coupling term that influences the creation of discontinuities in the vicinity of the significant locations. It is shown that the reconstruction corresponds to the use of a generic piecewise smoothness constraint. The model incorporates prior knowledge of discontinuity location into a reconstruction algorithm. The results appear favorable.<>