Reconstruction of Multiple Inhomogeneities for Circular Model in Electric Impedance Tomography

Electrical Impedance Tomography (EIT) is a method that shows the conductivity distribution usually from a cross-section within a geometry. In this study, it is aimed to visualize the inhomogenious distributions within the circle geometry. In this framework, the structure in EIT is handled as solution of forward and inverse problem. The forward problem of EIT is solved with the Finite Element Method (FEM) and necessary scenarios of the visual patterns referring different distribution positions and conductivities are generated for the use of inverse problem. The data set obtained with these scenarios were solved with Gauss-Newton (GN) reconstruction algorithm and selected Artificial Neural Networks (ANN) methods. The reconstruction images obtained by ANN and GN algorithm were compared over error criteria. As a result of this comparison, it has been observed that especially dual inhomogenious distributions in reconstruction phase can be obtained with less error performance with Radial Basis Function Neural Networks (RBFNN).