AN ITERATIVE METHOD FOR IMAGE RECONSTRUCTION FROM PROJECTIONS

Image reconstruction from projections gives rise to large ill-conditioned linear systems of equations. The large size of the systems often makes solution by an iterative method attractive. The sensitivity of the solution to perturbations in the system of equations generally makes it necessary to regularize the system before solution. We describe how Tikhonov regularization and Richardson iteration can be combined to yield an efficient adaptive method that determines both a suitable regularization parameter and a solution of the associated regularized system of equations.