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

TL;DR

This paper investigates active learning strategies to reduce annotation costs in training semantically-aware Neural Radiance Fields, demonstrating over 2X cost reduction through geometry-informed sample selection.

Contribution

It introduces a novel active learning approach that incorporates 3D geometric constraints for efficient sample selection in NeRF training.

Findings

Active learning reduces annotation costs by over 2X.

Geometry-aware sample selection improves training efficiency.

Proposed strategies outperform random sampling in NeRF annotation.

Abstract

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.