Computer Science > Computer Vision and Pattern Recognition
[Submitted on 4 Mar 2021 (this version), latest version 25 Jun 2021 (v4)]
Title:DONeRF: Towards Real-Time Rendering of Neural Radiance Fields using Depth Oracle Networks
View PDFAbstract:The recent research explosion around Neural Radiance Fields (NeRFs) shows that there is immense potential for implicitly storing scene and lighting information in neural networks, e.g., for novel view generation. However, one major limitation preventing the widespread use of NeRFs is the prohibitive computational cost of excessive network evaluations along each view ray, requiring dozens of petaFLOPS when aiming for real-time rendering on current devices. We show that the number of samples required for each view ray can be significantly reduced when local samples are placed around surfaces in the scene. To this end, we propose a depth oracle network, which predicts ray sample locations for each view ray with a single network evaluation. We show that using a classification network around logarithmically discretized and spherically warped depth values is essential to encode surface locations rather than directly estimating depth. The combination of these techniques leads to DONeRF, a dual network design with a depth oracle network as a first step and a locally sampled shading network for ray accumulation. With our design, we reduce the inference costs by up to 48x compared to NeRF. Using an off-the-shelf inference API in combination with simple compute kernels, we are the first to render raymarching-based neural representations at interactive frame rates (15 frames per second at 800x800) on a single GPU. At the same time, since we focus on the important parts of the scene around surfaces, we achieve equal or better quality compared to NeRF.
Submission history
From: Thomas Neff [view email][v1] Thu, 4 Mar 2021 18:55:09 UTC (45,522 KB)
[v2] Thu, 11 Mar 2021 18:57:56 UTC (45,517 KB)
[v3] Tue, 11 May 2021 09:56:38 UTC (45,517 KB)
[v4] Fri, 25 Jun 2021 09:05:10 UTC (25,758 KB)
References & Citations
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
Connected Papers (What is Connected Papers?)
CORE Recommender (What is CORE?)
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.