BakedSDF: Meshing Neural SDFs for Real-Time View Synthesis
SIGGRAPH 2023
-
Lior Yariv*
Weizmann Institute of Science
Google Research -
Peter Hedman*
Google Research
-
Christian Reiser
Tübingen AI Center
Google Research -
Dor Verbin
Google Research
-
Pratul P. Srinivasan
Google Research -
Richard Szeliski
Google Research -
Jonathan T. Barron
Google Research -
Ben Mildenhall
Google Research
Abstract
We present a method for reconstructing high-quality meshes of large unbounded real-world scenes suitable for photorealistic novel view synthesis. We first optimize a hybrid neural volume-surface scene representation designed to have well-behaved level sets that correspond to surfaces in the scene. We then bake this representation into a high-quality triangle mesh, which we equip with a simple and fast view-dependent appearance model based on spherical Gaussians. Finally, we optimize this baked representation to best reproduce the captured viewpoints, resulting in a model that can leverage accelerated polygon rasterization pipelines for real-time view synthesis on commodity hardware. Our approach outperforms previous scene representations for real-time rendering in terms of accuracy, speed, and power consumption, and produces high quality meshes that enable applications such as appearance editing and physical simulation.
Video
Real-Time Interactive Viewer Demos
Real Captured Scenes
BakedSDF
Our method bakes a hybrid VolSDF/mip-NeRF 360 scene representation into a triangle mesh, then optimizes a lightweight view-dependent appearance model to reproduce the training images. The final color of the rendered mesh is a sum of spherical Gaussian lobes (queried with the outgoing view direction) and a diffuse color. These parameters are stored per-vertex on the underlying mesh.
Mesh Extraction and Rendering
Citation
Acknowledgements
The website template was borrowed from Michaël Gharbi.