Fast and Memory-efficient Non-line-of-sight Imaging with Quasi-Fresnel Transform

2025-08-03

AI Summary

Existing non-line-of-sight (NLOS) imaging methods predominantly rely on 3D scene parameterizations, overlooking the intrinsic 2D structure of occluded objects—leading to high computational complexity and substantial memory overhead, thereby impeding real-time, high-resolution reconstruction on resource-constrained devices. This work proposes the first NLOS imaging framework based on 2D functional representation of hidden scenes. We introduce a Quasi-Fresnel transform to establish an exact forward model linking time-resolved measurements to the 2D scene geometry, and design an efficient inverse solver tailored to this formulation. Crucially, our approach explicitly exploits the problem’s inherent 2D nature, eliminating redundant 3D parameterization. Experiments demonstrate that, compared to state-of-the-art 3D methods, our framework reduces runtime and memory consumption by two to three orders of magnitude, while preserving reconstruction fidelity—enabling, for the first time, real-time, high-resolution NLOS imaging on mobile and embedded platforms.

Technology Category

• Computer Vision: 3D Computer Vision
• Search and Optimization: Non-convex Optimization
• Intelligent Robots: State Estimation

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Experiences and lessons learnt from Web-based algorithms and system deployments
• Security and Privacy: Data transparency and provenance
• Graph Algorithms and Modeling for the Web: Algorithms and analysis for incomplete, noisy, or partially observed Web-related graphs

Abstract

Non-line-of-sight (NLOS) imaging seeks to reconstruct hidden objects by analyzing reflections from intermediary surfaces. Existing methods typically model both the measurement data and the hidden scene in three dimensions, overlooking the inherently two-dimensional nature of most hidden objects. This oversight leads to high computational costs and substantial memory consumption, limiting practical applications and making real-time, high-resolution NLOS imaging on lightweight devices challenging. In this paper, we introduce a novel approach that represents the hidden scene using two-dimensional functions and employs a Quasi-Fresnel transform to establish a direct inversion formula between the measurement data and the hidden scene. This transformation leverages the two-dimensional characteristics of the problem to significantly reduce computational complexity and memory requirements. Our algorithm efficiently performs fast transformations between these two-dimensional aggregated data, enabling rapid reconstruction of hidden objects with minimal memory usage. Compared to existing methods, our approach reduces runtime and memory demands by several orders of magnitude while maintaining imaging quality. The substantial reduction in memory usage not only enhances computational efficiency but also enables NLOS imaging on lightweight devices such as mobile and embedded systems. We anticipate that this method will facilitate real-time, high-resolution NLOS imaging and broaden its applicability across a wider range of platforms.

Problem

• Reducing computational costs in NLOS imaging
• Minimizing memory usage for lightweight devices
• Enabling real-time high-resolution NLOS imaging

Innovation

• Uses 2D functions for hidden scene representation
• Applies Quasi-Fresnel transform for direct inversion
• Reduces computational complexity and memory usage