Three-dimensional Bayesian reconstruction applied to cardiac SPECT

摘要

The conventional maximum a posteriori probability estimators for image reconstruction in emission tomography suffer from oversmoothing of boundaries. To overcome this drawback, the authors propose a fully 3-D Bayesian reconstruction algorithm based on the 3-D compound Gauss-Markov random field model. Furthermore, they apply this algorithm to cardiac SPECT data to demonstrate its validity.