RoGUENeRF - A Robust Geometry-Consistent Universal Enhancer for NeRF
ECCV 2024
Sibi Catley-Chandar^† Richard Shaw^ Gregory Slabaugh† Eduardo Pérez-Pellitero^
^Huawei Noah's Ark Lab, †Queen Mary University Of London
Method
RoGUENeRF is a universal NeRF enhancer which leverages geometry-aware 3D alignment and 2D refinement to enhance any NeRF model while remaining view-consistent and being robust to inaccurate camera poses.
Results
If videos do not load, please try using Mozilla Firefox or download the video from the viewer.
MipNeRF360 - Treehill
Nerfacto - Garden
NeRF - Trex
View Consistency
Zoomed In Results - MipNeRF360 + Ours
We show zoomed in videos of our method below to demonstrate the view consistency of our approach.
Image Comparisons
Nerfacto - Room
NeRFLiX - Room
MipNeRF360 - Treehill
NeRFLiX - Treehill
Large Gaussian Camera Noise - Nerfacto
Large Gaussian Camera Noise - NeRFLiX
Medium Gaussian Camera Noise - Nerfacto
Medium Gaussian Camera Noise - NeRFLiX
@inproceedings{catleychandar2024roguenerf,
author={Sibi Catley-Chandar and Richard Shaw and Gregory Slabaugh and Eduardo Perez-Pellitero},
title={RoGUENeRF: A Robust Geometry-Consistent Universal Enhancer for NeRF},
booktitle=ECCV,
year={2024}
}
@article{catleychandar2024roguenerf,
title={RoGUENeRF: A Robust Geometry-Consistent Universal Enhancer for NeRF},
author={Catley-Chandar, Sibi and Shaw, Richard and Slabaugh, Gregory and Perez-Pellitero, Eduardo},
journal={arXiv preprint arXiv:2403.11909},
year={2024}
}