The use of neural network approximation models to speed up the optimisation process in electrical impedance tomography

A reduced approximation model technique based on neural networks is developed in order to increase the rate of convergence of an evolution strategy (ES) used for solving a non-destructive evaluation problem. The inverse problem investigated consists of identifying the geometry of discontinuities in a conductive material from Cauchy data measurements taken on the boundary. In this study, we use neural network (NN) approximation models in order to increase the rate of convergence of the optimisation algorithm and to efficiently detect, from a computational time point of view a subsurface cavity, such as a circle. The algorithm developed by combining evolution strategies and neural networks is found to be a robust, fast and efficient method for detecting the size and location of subsurface cavities.