Electromagnetic inverse-scattering solutions for inhomogeneous cross-sections of cylindrical biological bodies by a modified Born approximation and a stochastic Markov random fields approach

Abstract

Two novel methods of microwave imaging for biomedical applications are presented and discussed. The first is a deterministic method based on a modified Born approximation; it extends the validity of the usual Born approximation to strongly scattering bodies. The second is a stochastic method that leads to a Bayesian solution to the nonlinear problem of inverse scattering by modelling the dielectric properties of a scatterer and the electric field as suitable Markov random fields. Both methods exhibit the major feature of easily exploiting a priori knowledge, which can be available in many biomedical applications. Numerical results are reported to show the effectivenesses of the two methods.