Neural LightRig: Unlocking Accurate Object Normal and Material Estimation with Multi-Light Diffusion

TL;DR

Neural LightRig introduces a framework that uses multi-light diffusion priors from diffusion models to improve the accuracy of object normal and material estimation from a single image, enabling vivid relighting effects.

Contribution

It leverages large-scale diffusion models to generate multi-lighting conditions, reducing estimation uncertainty and significantly improving surface normal and material predictions.

Findings

Outperforms state-of-the-art methods in normal and material estimation

Enables vivid relighting effects from single images

Uses diffusion priors to generate consistent multi-light images

Abstract

Recovering the geometry and materials of objects from a single image is challenging due to its under-constrained nature. In this paper, we present Neural LightRig, a novel framework that boosts intrinsic estimation by leveraging auxiliary multi-lighting conditions from 2D diffusion priors. Specifically, 1) we first leverage illumination priors from large-scale diffusion models to build our multi-light diffusion model on a synthetic relighting dataset with dedicated designs. This diffusion model generates multiple consistent images, each illuminated by point light sources in different directions. 2) By using these varied lighting images to reduce estimation uncertainty, we train a large G-buffer model with a U-Net backbone to accurately predict surface normals and materials. Extensive experiments validate that our approach significantly outperforms state-of-the-art methods, enabling accurate surface normal and PBR material estimation with vivid relighting effects. Code and dataset are available on our project page at https://projects.zxhezexin.com/neural-lightrig.