Unsupervised Bayesian reconstruction of microwave images from real data

We address the problem of non-linear microwave imaging for the reconstruction of dielectric profiles. We propose a statistical based inversion algorithm, adopting the Bayesian (MAP) framework and a complex Gaussian MRF model for the image. The use of statistical algorithms for the estimation of the complex MRF parameter leads to a robust and effective non-linear inversion method. Some experiments on real data are able to show the good performance of the method.