Contrast-Source-Based Physics-Driven Neural Network for Inverse Scattering Problems

TL;DR

This paper introduces CSPDNN, a physics-driven neural network that efficiently solves inverse scattering problems by predicting induced currents and using adaptive regularization, validated through simulations and experiments.

Contribution

It proposes a contrast-source-based neural network that enhances efficiency and robustness in inverse scattering imaging, addressing limitations of existing UNNs.

Findings

Improved imaging accuracy demonstrated in simulations.

Enhanced robustness under noise and contrast variations.

Faster inference compared to traditional UNNs.

Abstract

Deep neural networks (DNNs) have recently been applied to inverse scattering problems (ISPs) due to their strong nonlinear mapping capabilities. However, supervised DNN solvers require large-scale datasets, which limits their generalization in practical applications. Untrained neural networks (UNNs) address this issue by updating weights from measured electric fields and prior physical knowledge, but existing UNN solvers suffer from long inference time. To overcome these limitations, this paper proposes a contrast-source-based physics-driven neural network (CSPDNN), which predicts the induced current distribution to improve efficiency and incorporates an adaptive total variation loss for robust reconstruction under varying contrast and noise conditions. The improved imaging performance is validated through comprehensive numerical simulations and experimental data.