Linearized Electrical Impedance Tomography

We address the problem of Electrical Impedance Tomography (EIT) as this is encountered in various applications of near surface geophysics. Based on a closed form integral model that relates the boundary observations to the electrical properties of the domain we suggest some extensions to the conventional `linearization--regularization' modeling approaches. Instead of truncating the Taylor series expansion of the forward operator to first-order accuracy, our methodology accounts for the linearization error as a multivariate random variable and uses stochastic simulation to obtain its statistical information.