Noise-adapted Neural Operator for Robust Non-Line-of-Sight Imaging

2025-08-13

AI Summary

In non-line-of-sight (NLOS) imaging, the weak and noise-prone indirect light signals severely degrade the accuracy and robustness of 3D reconstruction. To address this, we propose an end-to-end neural operator learning framework. First, we parameterize a neural operator to approximate the physical inverse mapping, integrated with differentiable light-transport modeling. Second, we introduce a noise-adaptive estimation module to dynamically respond to varying noise levels. Third, we design a spatiotemporal global–local distribution feature fusion mechanism to enhance fine-detail recovery. Leveraging deep algorithm unrolling, our method ensures efficient inference while significantly improving reconstruction accuracy and stability under sparse sampling and fast-scanning conditions—both in simulation and real-world experiments. This work establishes a new paradigm for practical NLOS imaging.

Technology Category

• Search and Optimization: Learning to Search
• Computer Vision: 3D Computer Vision
• Intelligent Robots: State Estimation

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Algorithms and analysis for incomplete, noisy, or partially observed Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Experiences and lessons learnt from Web-based algorithms and system deployments

Abstract

Computational imaging, especially non-line-of-sight (NLOS) imaging, the extraction of information from obscured or hidden scenes is achieved through the utilization of indirect light signals resulting from multiple reflections or scattering. The inherently weak nature of these signals, coupled with their susceptibility to noise, necessitates the integration of physical processes to ensure accurate reconstruction. This paper presents a parameterized inverse problem framework tailored for large-scale linear problems in 3D imaging reconstruction. Initially, a noise estimation module is employed to adaptively assess the noise levels present in transient data. Subsequently, a parameterized neural operator is developed to approximate the inverse mapping, facilitating end-to-end rapid image reconstruction. Our 3D image reconstruction framework, grounded in operator learning, is constructed through deep algorithm unfolding, which not only provides commendable model interpretability but also enables dynamic adaptation to varying noise levels in the acquired data, thereby ensuring consistently robust and accurate reconstruction outcomes. Furthermore, we introduce a novel method for the fusion of global and local spatiotemporal data features. By integrating structural and detailed information, this method significantly enhances both accuracy and robustness. Comprehensive numerical experiments conducted on both simulated and real datasets substantiate the efficacy of the proposed method. It demonstrates remarkable performance with fast scanning data and sparse illumination point data, offering a viable solution for NLOS imaging in complex scenarios.

Problem

• Robust NLOS imaging under noisy signal conditions
• Inverse mapping for 3D reconstruction from weak indirect light
• Dynamic noise adaptation in neural operator-based reconstruction

Innovation

• Noise-adapted neural operator for robust imaging
• Deep algorithm unfolding for interpretable reconstruction
• Global-local spatiotemporal feature fusion for accuracy