Deep Learning based Phaseless SAR without Born Approximation

In this paper, we present a phase retrieval approach from intensity measurements using a Deep Learning (DL) based Wirtinger Flow (WF) algorithm for the case where the measurement model is non-linear, and this non-linearity depends on the unknown signal. In the context of synthetic aperture radar (SAR), this is relevant to the image reconstruction problem for the scenario where the Born approximati...