Free-viewpoint Indoor Neural Relighting from Multi-view Stereo

TL;DR

This paper presents a neural relighting method for indoor scenes that enables interactive free-viewpoint navigation and realistic illumination changes, including shadows and glossy reflections, using multi-view stereo data and a learned implicit scene representation.

Contribution

It introduces a neural relighting approach that combines multi-view stereo reconstruction with a novel network for scene material and illumination modeling, enabling realistic relighting and free-viewpoint navigation.

Findings

Successfully relights real indoor scenes with complex glossy reflections

Enables interactive free-viewpoint navigation in relit scenes

Achieves coherent rendering of shadows and materials under new lighting

Abstract

We introduce a neural relighting algorithm for captured indoors scenes, that allows interactive free-viewpoint navigation. Our method allows illumination to be changed synthetically, while coherently rendering cast shadows and complex glossy materials. We start with multiple images of the scene and a 3D mesh obtained by multi-view stereo (MVS) reconstruction. We assume that lighting is well-explained as the sum of a view-independent diffuse component and a view-dependent glossy term concentrated around the mirror reflection direction. We design a convolutional network around input feature maps that facilitate learning of an implicit representation of scene materials and illumination, enabling both relighting and free-viewpoint navigation. We generate these input maps by exploiting the best elements of both image-based and physically-based rendering. We sample the input views to estimate diffuse scene irradiance, and compute the new illumination caused by user-specified light sources using path tracing. To facilitate the network's understanding of materials and synthesize plausible glossy reflections, we reproject the views and compute mirror images. We train the network on a synthetic dataset where each scene is also reconstructed with MVS. We show results of our algorithm relighting real indoor scenes and performing free-viewpoint navigation with complex and realistic glossy reflections, which so far remained out of reach for view-synthesis techniques.