Neuro-computational approaches to the FEM based electrical impedance tomography problems

Electrical impedance tomography (EIT) is a noninvasive technique to estimate the conductivity distribution inside the object. In EIT, driving currents are injected through the object and voltages are measured at the electrodes on the surface. Algorithms to estimate the conductivity distribution from the measured voltages are called reconstruction algorithms on which we propose neuro-computational ones in this paper. Our algorithms are based on the finite element method (FEM) to model the objects´ electric field, and conductivity distribution is expressed as a conductivity value of each element. In the first approach, we organized a layered network to describe the equation derived from the FEM. Using the network inversion technique, unknown conductivities are estimated. Second approach is a kind of random descent algorithm. A few homogeneous regions, consisting of a number of the elements, are modified when an assumed change decreases the error between the measured voltages and the calculated ones. We shows some numerical simulations to indicate the validity of these method