IMPEDANCE IMAGE RECONSTRUCTION WITH ARTIFICIAL NEURAL NETWORK IN ELECTRICAL IMPEDANCE TOMOGRAPHY

Abstract

Electrical impedance tomography views the electrical properties of theobjects by injecting current with surface electrodes and measuring voltages.Then using a reconstructing algorithm, from the measured voltage-currentvalues, conductivity distribution of the object calculated. Finding internalconductivity from surface voltage-current measurements is a reverse and illposedproblem.Therefore, high error sensitivity, and making approximations in conceivingcomplex computations cause to limited spatial resolution. The classic iterativeimage reconstruction algorithms have reconstruction errors. Accordingly,Electrical impedance tomography images suffer low accuracy. It is necessaryto evaluate the collected data from the object surface with a new approach. Inthis paper, the forward problem solved with the finite element method toreconstruct the conductivity distribution inside the object, the reverseproblem solved by the neural network approach. Image reconstruction speed,conceptual simplicity, and ease of implementation maintained by thisapproach.