The reconstruction from sparse- or few-view projections is one of important problems in computed tomography limited by the availability or feasibility of a large number of projections. Working with a small number of projections provides a lower radiation dose and a fast scan time, however an error associated with the sparse-view reconstruction increases significantly as the space sparsity increases that may cause the reconstruction process to diverge. The iterative reconstruction-re-projection (IRR) algorithm which uses filtered back projection (FBP) reconstruction has been used for the sparse-view computed tomography applications for several years. The IRR-TV method has been developed as a higher performance alternative to the IRR method by adding the total variation (TV) minimization. Here, we propose an algebraic iterative reconstruction-re-projection (AIRR) algorithm with the shearlet regularization. The AIRR coupled with the shearlet regularization in image space attains a better estimation in the projection space and yielded a higher performance based on subjective and objective quality metrics.
Digital Object Identifier (DOI)
Pour Yazdanpanah, Ali; E. Regentova, Emma; and Bebis, George
"Algebraic Iterative Reconstruction-Reprojection (AIRR) Method for High Performance Sparse-View CT Reconstruction,"
Applied Mathematics & Information Sciences: Vol. 10:
6, Article 2.
Available at: https://digitalcommons.aaru.edu.jo/amis/vol10/iss6/2