RelationField: Relate Anything in Radiance Fields

TL;DR

RelationField introduces a novel approach to extract and understand semantic relationships between objects within neural radiance fields, enabling open-vocabulary 3D scene graph generation and relationship-guided segmentation.

Contribution

It is the first method to directly extract inter-object relationships from neural radiance fields using implicit relationship queries and knowledge distillation from multi-modal LLMs.

Findings

Achieves state-of-the-art results in open-vocabulary 3D scene graph generation.

Excels in relationship-guided instance segmentation tasks.

Effectively models complex semantic relationships in 3D scenes.

Abstract

Neural radiance fields are an emerging 3D scene representation and recently even been extended to learn features for scene understanding by distilling open-vocabulary features from vision-language models. However, current method primarily focus on object-centric representations, supporting object segmentation or detection, while understanding semantic relationships between objects remains largely unexplored. To address this gap, we propose RelationField, the first method to extract inter-object relationships directly from neural radiance fields. RelationField represents relationships between objects as pairs of rays within a neural radiance field, effectively extending its formulation to include implicit relationship queries. To teach RelationField complex, open-vocabulary relationships, relationship knowledge is distilled from multi-modal LLMs. To evaluate RelationField, we solve open-vocabulary 3D scene graph generation tasks and relationship-guided instance segmentation, achieving state-of-the-art performance in both tasks. See the project website at https://relationfield.github.io.