ScoreField: Neural Inverse Scattering with Score-Based Generative Priors

2026-08-03

AI Summary

This work addresses the challenges of nonlinear full-wave constraints and prior modeling in electromagnetic inverse scattering by proposing a coupled implicit neural representation (INR) framework that separately parameterizes the dielectric contrast and induced current fields, jointly optimized under the Lippmann-Schwinger equation. Innovatively, a pretrained score-based diffusion model is incorporated as a contrast prior, with its gradient propagated back to the contrast INR via the chain rule—marking the first integration of score-based generative priors with coupled INRs in inverse scattering. The method significantly enhances reconstruction quality under strong multiple scattering conditions, outperforming existing full-wave and deep learning approaches on both synthetic and real Fresnel datasets, achieving an average PSNR gain of 1.8 dB on real data while effectively suppressing artifacts and improving fidelity.

Abstract

Designing an effective electromagnetic inverse-scattering solver requires faithful enforcement of nonlinear full-wave physics together with an expressive prior on the unknown permittivity contrast. We propose ScoreField, a neural inverse scattering framework that integrates coupled implicit neural representations (INRs) with a pretrained score-based generative prior. ScoreField employs two INRs to parameterize the permittivity contrast and the induced current fields, and jointly optimize them under the Lippmann-Schwinger equations. In addition to the implicit regularization by the INR architecture, the score model provides a learned prior gradient on the contrast, which is propagated to the contrast INR through the chain rule. This formulation enables ScoreField to effectively handle strong multiple scattering, where nonlinear wave interactions require accurate modeling of the coupled full-wave physics. We evaluate ScoreField on simulated weak- and strong-scattering benchmarks, the canonical Austria phantom, and experimental Fresnel measurements. We note that ScoreField significantly improves reconstruction fidelity and suppresses artifacts relative to classical full-wave methods and deep learning baselines, achieving an average PSNR improvement of $1.8 \, \mathrm{dB}$ over the best competing method on real Fresnel data.

Problem

• inverse scattering
• full-wave physics
• permittivity contrast
• multiple scattering
• generative prior

Innovation

• score-based generative model
• implicit neural representation
• inverse scattering
• full-wave physics
• Lippmann-Schwinger equation