MU-GeNeRF: Multi-view Uncertainty-guided Generalizable Neural Radiance Fields for Distractor-aware Scene

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Mu, Wenjie, Li, Zhan, Su, Chuanzhou, Shen, Xuanyi, Liu, Ziniu, Lu, Fan, Mo, Yujian, Zhao, Junqiao, Feng, Tiantian, Ye, Chen, Chen, Guang
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908979873972224
author Mu, Wenjie
Li, Zhan
Su, Chuanzhou
Shen, Xuanyi
Liu, Ziniu
Lu, Fan
Mo, Yujian
Zhao, Junqiao
Feng, Tiantian
Ye, Chen
Chen, Guang
author_facet Mu, Wenjie
Li, Zhan
Su, Chuanzhou
Shen, Xuanyi
Liu, Ziniu
Lu, Fan
Mo, Yujian
Zhao, Junqiao
Feng, Tiantian
Ye, Chen
Chen, Guang
contents Generalizable Neural Radiance Fields (GeNeRFs) enable high-quality scene reconstruction from sparse views and can generalize to unseen scenes. However, in real-world settings, transient distractors break cross-view structural consistency, corrupting supervision and degrading reconstruction quality. Existing distractor-free NeRF methods rely on per-scene optimization and estimate uncertainty from per-view reconstruction errors, which are not reliable for GeNeRFs and often misjudge inconsistent static structures as distractors. To this end, we propose MU-GeNeRF, a Multi-view Uncertainty-guided distractor-aware GeNeRF framework designed to alleviate GeNeRF's robust modeling challenges in the presence of transient distractions. We decompose distractor awareness into two complementary uncertainty components: Source-view Uncertainty, which captures structural discrepancies across source views caused by viewpoint changes or dynamic factors; and Target-view Uncertainty, which detects observation anomalies in the target image induced by transient distractors.These two uncertainties address distinct error sources and are combined through a heteroscedastic reconstruction loss, which guides the model to adaptively modulate supervision, enabling more robust distractor suppression and geometric modeling.Extensive experiments show that our method not only surpasses existing GeNeRFs but also achieves performance comparable to scene-specific distractor-free NeRFs.
format Preprint
id arxiv_https___arxiv_org_abs_2604_17965
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle MU-GeNeRF: Multi-view Uncertainty-guided Generalizable Neural Radiance Fields for Distractor-aware Scene
Mu, Wenjie
Li, Zhan
Su, Chuanzhou
Shen, Xuanyi
Liu, Ziniu
Lu, Fan
Mo, Yujian
Zhao, Junqiao
Feng, Tiantian
Ye, Chen
Chen, Guang
Computer Vision and Pattern Recognition
Generalizable Neural Radiance Fields (GeNeRFs) enable high-quality scene reconstruction from sparse views and can generalize to unseen scenes. However, in real-world settings, transient distractors break cross-view structural consistency, corrupting supervision and degrading reconstruction quality. Existing distractor-free NeRF methods rely on per-scene optimization and estimate uncertainty from per-view reconstruction errors, which are not reliable for GeNeRFs and often misjudge inconsistent static structures as distractors. To this end, we propose MU-GeNeRF, a Multi-view Uncertainty-guided distractor-aware GeNeRF framework designed to alleviate GeNeRF's robust modeling challenges in the presence of transient distractions. We decompose distractor awareness into two complementary uncertainty components: Source-view Uncertainty, which captures structural discrepancies across source views caused by viewpoint changes or dynamic factors; and Target-view Uncertainty, which detects observation anomalies in the target image induced by transient distractors.These two uncertainties address distinct error sources and are combined through a heteroscedastic reconstruction loss, which guides the model to adaptively modulate supervision, enabling more robust distractor suppression and geometric modeling.Extensive experiments show that our method not only surpasses existing GeNeRFs but also achieves performance comparable to scene-specific distractor-free NeRFs.
title MU-GeNeRF: Multi-view Uncertainty-guided Generalizable Neural Radiance Fields for Distractor-aware Scene
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2604.17965