Sources of Uncertainty in 3D Scene Reconstruction

Marcus Klasson, Riccardo Mereu, Juho Kannala, Arno Solin
9/10/2024
cs.CVcs.LG

No official code found. Best implementation: AaltoML/uncertainty-nerf-gs (30 stars, MIT, last commit 2024-12-10).

Abstract

The process of 3D scene reconstruction can be affected by numerous uncertainty sources in real-world scenes. While Neural Radiance Fields (NeRFs) and 3D Gaussian Splatting (GS) achieve high-fidelity rendering, they lack built-in mechanisms to directly address or quantify uncertainties arising from the presence of noise, occlusions, confounding outliers, and imprecise camera pose inputs. In this paper, we introduce a taxonomy that categorizes different sources of uncertainty inherent in these methods. Moreover, we extend NeRF- and GS-based methods with uncertainty estimation techniques, including learning uncertainty outputs and ensembles, and perform an empirical study to assess their ability to capture the sensitivity of the reconstruction. Our study highlights the need for addressing various uncertainty aspects when designing NeRF/GS-based methods for uncertainty-aware 3D reconstruction.

View on arXivView PDF

Code Implementations(1)

AaltoML/uncertainty-nerf-gsCommunity100%
MIT

Code release for the paper "Sources of Uncertainty in 3D Scene Reconstruction"

304PythonAug 28, 20241 years agoMIT

Similar Papers

Uncertainty-Aware 4D Gaussian Splatting for Monocular Occluded Human Rendering

85% match2026

Evaluating Modern Approaches in 3D Scene Reconstruction: NeRF vs Gaussian-Based Methods

84% match2024

RU4D-SLAM: Reweighting Uncertainty in Gaussian Splatting SLAM for 4D Scene Reconstruction

83% match2026

HERE: Hierarchical Active Exploration of Radiance Field with Epistemic Uncertainty Minimization

83% match2026

Uncertainty modeling for fine-tuned implicit functions

83% match2024

These Magic Moments: Differentiable Uncertainty Quantification of Radiance Field Models

83% match2025

Cite this paper

@article{klasson2024sources,
  title  = {Sources of Uncertainty in 3D Scene Reconstruction},
  author = {Marcus Klasson and Riccardo Mereu and Juho Kannala and Arno Solin},
  year   = {2024},
  eprint = {2409.06407},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2409.06407v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2409.06407.svg)](https://codeofpaper.com/paper/2409.06407)

Discussion