HaLo-NeRF: Learning Geometry-Guided Semantics for Exploring Unconstrained Photo Collections

TL;DR

HaLo-NeRF introduces a novel system that combines vision-and-language models with Internet data to localize and understand semantic regions in large-scale landmark scenes, enhancing 3D scene representations.

Contribution

It presents a new localization approach that integrates large-scale Internet data and vision-language models to improve semantic understanding in 3D landmark scenes.

Findings

Accurately localizes semantic concepts in landmark scenes

Outperforms existing 3D models and 2D segmentation baselines

Leverages Internet data for fine-grained semantic knowledge

Abstract

Internet image collections containing photos captured by crowds of photographers show promise for enabling digital exploration of large-scale tourist landmarks. However, prior works focus primarily on geometric reconstruction and visualization, neglecting the key role of language in providing a semantic interface for navigation and fine-grained understanding. In constrained 3D domains, recent methods have leveraged vision-and-language models as a strong prior of 2D visual semantics. While these models display an excellent understanding of broad visual semantics, they struggle with unconstrained photo collections depicting such tourist landmarks, as they lack expert knowledge of the architectural domain. In this work, we present a localization system that connects neural representations of scenes depicting large-scale landmarks with text describing a semantic region within the scene, by harnessing the power of SOTA vision-and-language models with adaptations for understanding landmark scene semantics. To bolster such models with fine-grained knowledge, we leverage large-scale Internet data containing images of similar landmarks along with weakly-related textual information. Our approach is built upon the premise that images physically grounded in space can provide a powerful supervision signal for localizing new concepts, whose semantics may be unlocked from Internet textual metadata with large language models. We use correspondences between views of scenes to bootstrap spatial understanding of these semantics, providing guidance for 3D-compatible segmentation that ultimately lifts to a volumetric scene representation. Our results show that HaLo-NeRF can accurately localize a variety of semantic concepts related to architectural landmarks, surpassing the results of other 3D models as well as strong 2D segmentation baselines. Our project page is at https://tau-vailab.github.io/HaLo-NeRF/.