Emergent Extreme-View Geometry in 3D Foundation Models

TL;DR

This paper reveals that 3D foundation models inherently understand extreme-view geometry and proposes a lightweight tuning method to enhance this ability, supported by a new challenging benchmark dataset.

Contribution

It uncovers emergent extreme-view geometric reasoning in 3DFMs and introduces a simple tuning scheme to improve pose estimation under extreme viewpoints.

Findings

3DFMs exhibit emergent understanding of extreme-view geometry

Lightweight bias tuning improves pose estimation without degrading depth quality

Introduces MegaUnScene benchmark for unseen Internet scenes

Abstract

3D foundation models (3DFMs) have recently transformed 3D vision, enabling joint prediction of depths, poses, and point maps directly from images. Yet their ability to reason under extreme, non-overlapping views remains largely unexplored. In this work, we study their internal representations and find that 3DFMs exhibit an emergent understanding of extreme-view geometry, despite never being trained for such conditions. To further enhance these capabilities, we introduce a lightweight alignment scheme that refines their internal 3D representation by tuning only a small subset of backbone bias terms, leaving all decoder heads frozen. This targeted adaptation substantially improves relative pose estimation under extreme viewpoints without degrading per-image depth or point quality. Additionally, we contribute MegaUnScene, a new benchmark of Internet scenes unseen by existing 3DFMs, with dedicated test splits for both relative pose estimation and dense 3D reconstruction. All code and data will be released.