A fast method for simultaneous reconstruction and segmentation in X-ray CT application

In this paper, we propose a new method to solve the minimization problem in simultaneous reconstruction and segmentation (SRS) model for X-ray computed tomography (CT). The SRS uses Bayes' rule maximum posteriori (MAP) estimate on hidden Markov measure field (HMMFM). original [Romanov M, Dahl AB, Dong Y, Hansen PC. Simultaneous tomographic with class priors. Inverse Problems Sci Eng. 2016;24(8):1432–1453] includes subproblem logarithmic-summation (log-sum) term, which is non-separable classification index. This was solved by Frank–Wolfe algorithm, very time consuming especially when dealing large-scale CT problems. starting point of paper commutativity log-sum operations, where could be transformed into sum-log introducing an auxiliary variable. corresponding separable. By applying primal-dual turns several easy-to-solve convex subproblems. addition, introduce improved adding Tikhonov regularization model, give some convergence results proposed methods. Experimental demonstrate that methods produce comparable compared much less CPU time.