Exploring Active Learning for Label-Efficient Training of Semantic Neural Radiance Field

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Zhu, Yuzhe, Cai, Lile, Lu, Kangkang, Liu, Fayao, Yang, Xulei
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912498189336576
author Zhu, Yuzhe
Cai, Lile
Lu, Kangkang
Liu, Fayao
Yang, Xulei
author_facet Zhu, Yuzhe
Cai, Lile
Lu, Kangkang
Liu, Fayao
Yang, Xulei
contents Neural Radiance Field (NeRF) models are implicit neural scene representation methods that offer unprecedented capabilities in novel view synthesis. Semantically-aware NeRFs not only capture the shape and radiance of a scene, but also encode semantic information of the scene. The training of semantically-aware NeRFs typically requires pixel-level class labels, which can be prohibitively expensive to collect. In this work, we explore active learning as a potential solution to alleviate the annotation burden. We investigate various design choices for active learning of semantically-aware NeRF, including selection granularity and selection strategies. We further propose a novel active learning strategy that takes into account 3D geometric constraints in sample selection. Our experiments demonstrate that active learning can effectively reduce the annotation cost of training semantically-aware NeRF, achieving more than 2X reduction in annotation cost compared to random sampling.
format Preprint
id arxiv_https___arxiv_org_abs_2507_17351
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Exploring Active Learning for Label-Efficient Training of Semantic Neural Radiance Field
Zhu, Yuzhe
Cai, Lile
Lu, Kangkang
Liu, Fayao
Yang, Xulei
Computer Vision and Pattern Recognition
Neural Radiance Field (NeRF) models are implicit neural scene representation methods that offer unprecedented capabilities in novel view synthesis. Semantically-aware NeRFs not only capture the shape and radiance of a scene, but also encode semantic information of the scene. The training of semantically-aware NeRFs typically requires pixel-level class labels, which can be prohibitively expensive to collect. In this work, we explore active learning as a potential solution to alleviate the annotation burden. We investigate various design choices for active learning of semantically-aware NeRF, including selection granularity and selection strategies. We further propose a novel active learning strategy that takes into account 3D geometric constraints in sample selection. Our experiments demonstrate that active learning can effectively reduce the annotation cost of training semantically-aware NeRF, achieving more than 2X reduction in annotation cost compared to random sampling.
title Exploring Active Learning for Label-Efficient Training of Semantic Neural Radiance Field
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2507.17351