KITTI-360

Method

Multi-view stereo regulated NeRF [MVSRegNeRF]


Submitted on 30 Apr. 2023 15:33 by
Bian FeiHu (East China Normal university)

Running time:2 s
Environment:1 core @ 2.5 Ghz (C/C++)

Method Description:
We leverage instantNGP with Multi-view stereo method
to
regulate the
underline geometry of NeRF, which can improve the
result
of novel view
synthesis task on large scale scene like kitti360
Parameters:
\alpha=0.2
Latex Bibtex:
@Article{mvsregnerf,
author={Bian, Feihu
and Xiong, Suya
and Yi, Ran
and Ma, Lizhuang},
title={Multi-view stereo-regulated NeRF for urban scene
novel view synthesis},
journal={The Visual Computer},
year={2024},
month={Mar},
day={27},
issn={1432-2315},
doi={10.1007/s00371-024-03321-6},
url={https://doi.org/10.1007/s00371-024-03321-6}
}

Detailed Results

This page provides detailed results for the method(s) selected. For the first 10 test images, we display the synthesized image and an error image. The error image visualizes the SSIM at every pixel. As the range of SSIM is within [-1,1] with 1 indicating the best performance and -1 indicating the worst, we visualize 1 - (1 + SSIM) / 2 such that bright region means large error and dark means low error.

Test Set Average

PSNR SSIM LPIPS
22.48 0.829 0.256
This table as LaTeX

Test Image 0

Prediction

Error


Test Image 1

Prediction

Error


Test Image 2

Prediction

Error


Test Image 3

Prediction

Error


Test Image 4

Prediction

Error


Test Image 5

Prediction

Error


Test Image 6

Prediction

Error


Test Image 7

Prediction

Error


Test Image 8

Prediction

Error


Test Image 9

Prediction

Error





eXTReMe Tracker