A Microwave Imaging Reconstruction Technique with Mild-Data-Driven Regularization in Variable-Exponent Lebesgue Spaces

A microwave imaging approach based on mild-data-driven inversion technique is proposed. In the developed approach, a-priori information regarding the class of targets under investigation is inserted through a data-driven regularization term inside a variable-exponent Lebesgue-space inversion procedure. The method has been preliminarily assessed for detecting and characterizing inclusions in partially known host targets, showing promising results compared with the standard inversion scheme.