Estimation of conductivity distribution based on fast inversion using eigenvalue and eigenvector in electrical impedance tomography

In order to solve the inverse solution for conductivity distribution in electrical impedance tomography, the one-step Gauss-Newton method is usually employed. Major computational time is involved in the calculation of inverse term of the Jacobian matrix and the complexity increases with the number of electrodes and finite elements. Therefore, to reduce the computational time, the inverse term is replaced with a summation term based on the eigenvalue and eigenvector in the inverse solver. In this paper, a fast inversion method using eigenvalue and eigenvector is developed to monitor the conductivity distribution. Therefore, using the proposed method the computation of inverse matrix is avoided resulting in decrease of the on-line computational time. Numerical simulations and experiments have been carried out to evaluate the performance of the proposed method.