Generalizable Non-Line-of-Sight Imaging with Learnable Physical Priors

TL;DR

This paper introduces a learning-based NLOS imaging method with learnable physical priors that generalizes well across real-world datasets, even with low SNR, by adaptively compensating paths and spectrum selection.

Contribution

It proposes a novel approach with Learnable Path Compensation and Adaptive Phasor Field, enabling better generalization and robustness in NLOS imaging compared to prior fixed-prior methods.

Findings

Effective in low SNR conditions

Generalizes across different real-world datasets

Trained solely on synthetic data

Abstract

Non-line-of-sight (NLOS) imaging, recovering the hidden volume from indirect reflections, has attracted increasing attention due to its potential applications. Despite promising results, existing NLOS reconstruction approaches are constrained by the reliance on empirical physical priors, e.g., single fixed path compensation. Moreover, these approaches still possess limited generalization ability, particularly when dealing with scenes at a low signal-to-noise ratio (SNR). To overcome the above problems, we introduce a novel learning-based solution, comprising two key designs: Learnable Path Compensation (LPC) and Adaptive Phasor Field (APF). The LPC applies tailored path compensation coefficients to adapt to different objects in the scene, effectively reducing light wave attenuation, especially in distant regions. Meanwhile, the APF learns the precise Gaussian window of the illumination function for the phasor field, dynamically selecting the relevant spectrum band of the transient measurement. Experimental validations demonstrate that our proposed approach, only trained on synthetic data, exhibits the capability to seamlessly generalize across various real-world datasets captured by different imaging systems and characterized by low SNRs.