Floaters No More: Radiance Field Gradient Scaling for Improved Near-Camera Training

Loading...
Thumbnail Image
Date
2023
Journal Title
Journal ISSN
Volume Title
Publisher
The Eurographics Association
Abstract
NeRF acquisition typically requires careful choice of near planes for the different cameras or suffers from background collapse, creating floating artifacts on the edges of the captured scene. The key insight of this work is that background collapse is caused by a higher density of samples in regions near cameras. As a result of this sampling imbalance, near-camera volumes receive significantly more gradients, leading to incorrect density buildup. We propose a gradient scaling approach to counter-balance this sampling imbalance, removing the need for near planes, while preventing background collapse. Our method can be implemented in a few lines, does not induce any significant overhead, and is compatible with most NeRF implementations.
Description

CCS Concepts: Computing methodologies -> Rendering

        
@inproceedings{
10.2312:sr.20231122
, booktitle = {
Eurographics Symposium on Rendering
}, editor = {
Ritschel, Tobias
and
Weidlich, Andrea
}, title = {{
Floaters No More: Radiance Field Gradient Scaling for Improved Near-Camera Training
}}, author = {
Philip, Julien
and
Deschaintre, Valentin
}, year = {
2023
}, publisher = {
The Eurographics Association
}, ISSN = {
1727-3463
}, ISBN = {
978-3-03868-229-5
978-3-03868-228-8
}, DOI = {
10.2312/sr.20231122
} }
Citation