Domain Reduction Strategy for Non-Line-of-Sight Imaging

Abstract

This paper presents a novel optimization-based method for non-line-of-sight (NLOS) imaging that aims to reconstruct hidden scenes under general setups with significantly reduced reconstruction time. In NLOS imaging, the visible surfaces of the target objects are notably sparse. To mitigate unnecessary computations arising from empty regions, we design our method to render the transients through partial propagations from a continuously sampled set of points from the hidden space. Our method is capable of accurately and efficiently modeling the view-dependent reflectance using surface normals, which enables us to obtain surface geometry as well as albedo. In this pipeline, we propose a novel domain reduction strategy to eliminate superfluous computations in empty regions. During the optimization process, our domain reduction procedure periodically prunes the empty regions from our sampling domain in a coarse-to-fine manner, leading to substantial improvement in efficiency. We demonstrate the effectiveness of our method in various NLOS scenarios with sparse scanning patterns. Experiments conducted on both synthetic and real-world data support the efficacy in general NLOS scenarios, and the improved efficiency of our method compared to the previous optimization-based solutions.

Overview

(a) A typical Non-line-of-sight (NLOS) imaging system. NLOS imaging aims to reconstruct scenes that are hidden in direct line-of-sight systems, with a laser illuminating a relay wall, and a time-resolved detector recording the returning photons. Non-line-of-sight (NLOS) imaging aims to reconstruct hidden scenes from measurements of indirect reflections.

(b) Considering practical applications of NLOS imaging, there are numerous scenarios where previous inverse NLOS methods are not sufficient to meet the demands. Our method is capable of reconstructing both albedo and surface normal of the hidden objects in general scenarios, including non-confocal, non-planar relay walls and sparse sampling.

(c) Our domain reduction procedure gradually prunes empty regions in a coarse-to-fine manner, achieving about 20× acceleration of the reconstruction time. The memory efficiency of our method enables reconstruction of high-resolution volumes with a single commercial GPU.

Method

(a) In our reconstruction scheme, we divide hidden space to a grid shape and assign albedo and surface normal for each vertex. Input points are randomly sampled from the hidden space. Then the light propagation from each point is computed, of which superposition is used as predicted transients. The variables are optimized by minimizing the L2 distance. Our domain is gradually reduced by pruning the empty regions during the optimization.

(b) We compute the forward light propagation function from each point. The arrival time at scan point is first identified, and then the fall-off term and the cosine term are computed using point-wise albedo and normal.

Results

Confocal 32×32 Results

We first evaluate our method on 32×32 measurements. As can be seen, our method achieves the highest performance.

Normal Maps

Compared to DLCT, our method delivers normal maps in high-quality, reconstructing details of the objects, with significantly reduced scanning time (32×32 measurements).

Results in Various Scenarios

(a) Results on the real-world 32×32 transients of the NT instance, measured with a non-planar relay wall. (b) Results on non-confocal ZNLOS Bunny. (c) Results on Bunny with various sampling resolutions. Our method consistently delivers high-quality results in various scanning scenarios.

BibTeX

@article{shim2024drs,
  author    = {Shim, Hyunbo and Cho, In and Daekyu, Kwon and Kim, Seon Joo},
  title     = {Domain Reduction Strategy for Non Line of Sight Imaging},
  journal   = {Proceedings of the European Conference on Computer Vision (ECCV)},
  year      = {2024},
}