The Use of the Approximation Error Method and Bayesian Inference to Introduce Anatomical and Physiological Prior Information into D-bar AlgorithmsMS20

In Electrical impedance tomography, it was shown that the use of prior information and the Approximation Error Method improves the spatial resolution using the Gauss-Newton method. However, the D-bar method’s approach to solve the inverse problem makes difficult to include statistical priors in the algorithm. In this work a method of correcting the data based on Bayesian techniques to incorporate statistical priors in the D-bar algorithm is proposed. Results are shown on experimental tank data.

This presentation is part of Minisymposium “MS20 - Advances in Reconstruction Methods for Electrical Impedance Tomography (3 parts)
organized by: Melody Alsaker (Gonzaga University) , Samuli Siltanen (University of Helsinki) .

Talles Santos (Polytechnic School of University of São Paulo)
bayesian methods, image reconstruction, inverse problems, statistical inverse estimation methods