Exploiting Multiple Priors for Neural 3D Indoor Reconstruction

TL;DR

This paper introduces a neural implicit modeling technique that combines multiple regularization strategies, including depth priors and self-supervised normal regularization, to improve large-scale indoor 3D reconstructions from images.

Contribution

It presents a novel method leveraging multiple priors and self-supervision to enhance neural 3D indoor scene reconstruction, addressing limitations of existing approaches.

Findings

Achieves state-of-the-art results on challenging indoor scenes.

Effectively uses depth priors to anchor and guide reconstruction.

Handles challenging lighting conditions with learnable exposure compensation.

Abstract

Neural implicit modeling permits to achieve impressive 3D reconstruction results on small objects, while it exhibits significant limitations in large indoor scenes. In this work, we propose a novel neural implicit modeling method that leverages multiple regularization strategies to achieve better reconstructions of large indoor environments, while relying only on images. A sparse but accurate depth prior is used to anchor the scene to the initial model. A dense but less accurate depth prior is also introduced, flexible enough to still let the model diverge from it to improve the estimated geometry. Then, a novel self-supervised strategy to regularize the estimated surface normals is presented. Finally, a learnable exposure compensation scheme permits to cope with challenging lighting conditions. Experimental results show that our approach produces state-of-the-art 3D reconstructions in challenging indoor scenarios.