Gauss-Markov-Potts Priors for Images in Computer Tomography Resulting to Joint Optimal Reconstruction and segmentation

摘要: In many applications of Computed Tomography (CT), we know that the object under test is composed a finite number materials meaning images to be reconstructed are homogeneous area. To account for this prior knowledge, propose family Gauss-Markov fields with hidden Potts label fields. Then, using these models in Bayesian inference framework, able jointly reconstruct image and segment it an optimal way. paper, first present models, then appropriate computational methods (MCMC or Variational Bayes) compute Joint Maximum A Posteriori (JMAP) posterior mean estimators. We finally provide few results showing efficiency proposed CT very limited angle projections.