Materialist: Physically Based Editing Using Single-Image Inverse Rendering

TL;DR

Materialist introduces a neural-initialized, physically based rendering pipeline for single-image inverse rendering, enabling accurate material editing, object insertion, relighting, and transparency editing without multi-view data.

Contribution

It presents a novel hybrid approach combining neural predictions with rigorous optimization for single-image inverse rendering, improving accuracy and versatility.

Findings

Outperforms existing methods on synthetic and real datasets.

Enables editing of material transparency via ray-traced refraction.

Achieves competitive envmap estimation performance.

Abstract

Achieving physically consistent image editing remains a significant challenge in computer vision. Existing image editing methods typically rely on neural networks, which struggle to accurately handle shadows and refractions. Conversely, physics-based inverse rendering often requires multi-view optimization, limiting its practicality in single-image scenarios. In this paper, we propose Materialist, a neural-initialized physically based rendering pipeline for single-image inverse rendering. Unlike previous hybrid methods that use physics to guide neural generation, our method leverages neural networks to predict initial material properties, which are then rigorously optimized via progressive differentiable rendering. Our approach enables a range of applications, including material editing, object insertion, and relighting, while also introducing an effective method for editing material transparency via ray-traced refraction without requiring full scene geometry. Furthermore, our envmap estimation method also achieves competitive performance, further enhancing the accuracy of image editing task. Experiments demonstrate strong performance across synthetic and real-world datasets, excelling even on challenging out-of-domain images.