Physically Controllable Relighting of Photographs

TL;DR

This paper introduces a self-supervised, physically based neural relighting method that allows users to control scene illumination in real-world photographs by combining traditional rendering with neural networks.

Contribution

It presents a novel pipeline that infers a 3D scene representation from monocular images and enables explicit physical control over lighting in in-the-wild photographs.

Findings

Achieves photorealistic relighting with user-controlled illumination.

Uses a differentiable rendering process for self-supervised training.

Bridges the gap between traditional graphics and neural rendering for relighting.

Abstract

We present a self-supervised approach to in-the-wild image relighting that enables fully controllable, physically based illumination editing. We achieve this by combining the physical accuracy of traditional rendering with the photorealistic appearance made possible by neural rendering. Our pipeline works by inferring a colored mesh representation of a given scene using monocular estimates of geometry and intrinsic components. This representation allows users to define their desired illumination configuration in 3D. The scene under the new lighting can then be rendered using a path-tracing engine. We send this approximate rendering of the scene through a feed-forward neural renderer to predict the final photorealistic relighting result. We develop a differentiable rendering process to reconstruct in-the-wild scene illumination, enabling self-supervised training of our neural renderer on raw image collections. Our method represents a significant step in bringing the explicit physical control over lights available in typical 3D computer graphics tools, such as Blender, to in-the-wild relighting.