A microwave object recognition approach based on neural networks

In this paper a neural network approach is applied to the solution of microwave inverse scattering problems. In particular, cylindrical objects buried in an unavailable domain are taken into account. Different sets of testing data are considered in order to verify the capability of the network to reconstruct the geometric and dielectric characteristics of an unknown target starting from the knowledge of the scattered field values. Both lossless homogeneous and lossy inhomogeneous domains are taken into account.