A convergence estimator for the iterative reconstruction of electrical impedance tomography images

The reconstruction of electrical impedance tomography (EIT) images poses an inverse problem that requires iterative solution. However, due to its severe ill-conditioning the EIT problem often fails to converge toward a physically meaningful minimum. In this paper, we present an estimator that tracks the convergence of an iterative EIT image reconstruction process by comparing the trajectories of the image conductivities and the objective function. In a computer simulation, this criterion was fulfilled for all reconstructions that converged to a meaningful image.