Multi-resolution compressive sensing inversion of scattering data

This paper proposes a novel technique for retrieving the dielectric features of weak scatterers in microwave imaging by means of a Compressive Sensing (CS)-based method enhanced by a multi-zoom strategy. A Relevance Vector Machine (RVM) is used to invert the data of the problem recast in a Bayesian framework, exploiting the combination of the a-priori information on the sparseness of the unknowns and the acquired knowledge during the iterative multi-scaling methodology. Representative results are presented to illustrate advantages and limitations of the proposed method.