VCURF: Virtual Camera-based Uncertainty of Radiance Fields

2026-10-02

AI Summary

This study addresses the lack of effective pixel-level uncertainty quantification methods for novel view synthesis using radiance fields by proposing the VCURF framework. This approach estimates rendering inconsistencies through virtual camera sampling, enabling black-box uncertainty estimation. Its core innovation lies in its compatibility with both implicit and explicit representations, such as NeRF and 3D Gaussian Splatting, without requiring any modifications to the underlying model architectures. Furthermore, it overcomes the reliance on specific viewpoint selections prevalent in existing literature. Extensive experiments across multiple datasets demonstrate that VCURF significantly outperforms current baseline methods in pixel-level uncertainty estimation tasks.

Abstract

Radiance fields, implemented with either implicit (NeRF) or explicit (Gaussian Splatting) representations, are advancing the state of the art in novel view synthesis at a rapid pace. Even though the rendered views they generate are often compelling, they are not free of errors. In this paper, we propose a new approach for pixel-wise uncertainty quantification based on measuring the inconsistencies among renderings by the radiance field model in virtual cameras sampled near the target viewpoint. We named our approach VCURF for Virtual Camera-based Uncertainty of Radiance Fields. VCURF treats the radiance field model as a black box, only assuming that it is capable of rendering color and depth on demand. This property makes our approach applicable to both NeRF and GS models without any modification. Our experiments on a combination of datasets, radiance field models and baselines demonstrate VCURF's effectiveness in pixel-wise uncertainty estimation. We conclude the paper with findings that question the way view selection is tackled by the majority of the current literature.

Problem

• Radiance Fields
• Uncertainty Quantification
• Novel View Synthesis
• Pixel-wise Uncertainty
• NeRF

Innovation

• Uncertainty Quantification
• Radiance Fields
• Virtual Camera
• Neural Radiance Fields (NeRF)
• Gaussian Splatting