Coordinate-Residual Physics-Driven Neural Network for Electromagnetic Inverse Scattering
2026-08-10 • Machine Learning
Machine Learning
AI summaryⓘ
The authors propose a new method called CRPDNN to improve 3-D electromagnetic imaging, which is usually hard because of noisy data and high computational costs. Unlike previous methods that need a rough initial guess, their approach directly reconstructs images using spatial coordinates and a neural network, leading to better accuracy and much faster results. They tested their method on synthetic and real data, showing it works well even with noise and is more stable than other physics-driven neural networks. Their code is also publicly available for others to use and explore.
electromagnetic inverse scattering3-D imagingphysics-driven neural networks (PDNN)coordinate-residual networksnonlinear inverse problemscontrast reconstructionnoise robustnesscomputational efficiencyCSI methodFresnel experiments
Authors
Yutong Du, Zicheng Liu, Bo Qi, Yali Zong, Peixian Han
Abstract
Electromagnetic inverse scattering is a nonlinear and ill-posed problem, where accurate reconstruction is challenging due to measurement limitations, noise, and high computational costs, especially for 3-D imaging. Although physics-driven neural networks (PDNNs) reduce the dependence on labeled training data, existing accelerated PDNN frameworks often rely on preliminary reconstruction-based region selection, which may introduce instability when the selected region is inaccurate. In this paper, a coordinate-residual physics-driven neural network (CRPDNN) is proposed for 3-D electromagnetic inverse scattering. The proposed solver directly reconstructs the unknown contrast distribution using normalized spatial coordinates and a residual convolutional network, without requiring a preliminary reconstruction. For the reported noise-free 3-D synthetic cases, CRPDNN achieves an average relative error of 2.10\%, compared with 7.97\% for CSI and 3.99\% for $L_{2/3}$-FBE-WCIE, while providing approximately 5.5- and 12.1-fold speedups over the two baselines, respectively. Supplementary 2-D comparisons further confirm its stability and computational efficiency relative to existing PDNN frameworks. CRPDNN also maintains reliable reconstruction performance under noisy measurements, and the 3-D Fresnel experiments further indicate its potential for practical imaging applications. The related code is available at https://github.com/Physics-driven-methods.