A model-aware inexact Newton scheme for electrical impedance tomography

This work gives new insights into the EIT model. Firstly, a novel relation between the conductivity and the data is derived, giving quantitative insights about the instability of the inverse problem. Secondly, a reconstruction framework is introduced which estimates unknown model parameters and then solves the problem with a tailored Newton method. Additional problem-specific optimizations are incorporated into the framework. Simulations verify its efficiency for simulated and measured data.