Neural Radiance Field Codebooks

TL;DR

Neural Radiance Field Codebooks (NRC) introduce a scalable, object-centric representation learning method for complex scenes, enabling better transfer to downstream tasks like navigation, segmentation, and depth ordering.

Contribution

NRC is a novel approach that learns object codes for scene reconstruction, discovering transferable visual and geometric patterns across scenes.

Findings

Outperforms existing methods in object navigation success rate by 3.1%.

Achieves 29% relative improvement in unsupervised segmentation.

Improves depth ordering accuracy by 5.5% in THOR.

Abstract

Compositional representations of the world are a promising step towards enabling high-level scene understanding and efficient transfer to downstream tasks. Learning such representations for complex scenes and tasks remains an open challenge. Towards this goal, we introduce Neural Radiance Field Codebooks (NRC), a scalable method for learning object-centric representations through novel view reconstruction. NRC learns to reconstruct scenes from novel views using a dictionary of object codes which are decoded through a volumetric renderer. This enables the discovery of reoccurring visual and geometric patterns across scenes which are transferable to downstream tasks. We show that NRC representations transfer well to object navigation in THOR, outperforming 2D and 3D representation learning methods by 3.1% success rate. We demonstrate that our approach is able to perform unsupervised segmentation for more complex synthetic (THOR) and real scenes (NYU Depth) better than prior methods (29% relative improvement). Finally, we show that NRC improves on the task of depth ordering by 5.5% accuracy in THOR.