MaterialFusion: Enhancing Inverse Rendering with Material Diffusion Priors

TL;DR

MaterialFusion introduces a novel inverse rendering approach that leverages a 2D diffusion prior to improve the accuracy of material and albedo estimation, resulting in better relighting under new lighting conditions.

Contribution

We propose MaterialFusion, integrating a 2D diffusion model prior with inverse rendering to enhance material and albedo recovery for more accurate relighting.

Findings

Significantly improves relighting accuracy on synthetic and real datasets.

Uses StableMaterial diffusion prior trained on BlenderVault dataset.

Outperforms previous methods in relighting under diverse illumination.

Abstract

Recent works in inverse rendering have shown promise in using multi-view images of an object to recover shape, albedo, and materials. However, the recovered components often fail to render accurately under new lighting conditions due to the intrinsic challenge of disentangling albedo and material properties from input images. To address this challenge, we introduce MaterialFusion, an enhanced conventional 3D inverse rendering pipeline that incorporates a 2D prior on texture and material properties. We present StableMaterial, a 2D diffusion model prior that refines multi-lit data to estimate the most likely albedo and material from given input appearances. This model is trained on albedo, material, and relit image data derived from a curated dataset of approximately ~12K artist-designed synthetic Blender objects called BlenderVault. we incorporate this diffusion prior with an inverse rendering framework where we use score distillation sampling (SDS) to guide the optimization of the albedo and materials, improving relighting performance in comparison with previous work. We validate MaterialFusion's relighting performance on 4 datasets of synthetic and real objects under diverse illumination conditions, showing our diffusion-aided approach significantly improves the appearance of reconstructed objects under novel lighting conditions. We intend to publicly release our BlenderVault dataset to support further research in this field.