FunRec: Reconstructing Functional 3D Scenes from Egocentric Interaction Videos

TL;DR

FunRec is a novel method for reconstructing functional 3D indoor scenes from egocentric RGB-D videos, enabling accurate, interaction-aware digital twins without controlled setups.

Contribution

It introduces a new approach that directly reconstructs articulated, interaction-aware 3D scenes from in-the-wild videos, surpassing prior methods in accuracy and applicability.

Findings

Achieves up to +50 mIoU improvement in part segmentation.

5-10 times lower articulation and pose errors.

Significantly higher reconstruction accuracy on benchmarks.

Abstract

We present FunRec, a method for reconstructing functional 3D digital twins of indoor scenes directly from egocentric RGB-D interaction videos. Unlike existing methods on articulated reconstruction, which rely on controlled setups, multi-state captures, or CAD priors, FunRec operates directly on in-the-wild human interaction sequences to recover interactable 3D scenes. It automatically discovers articulated parts, estimates their kinematic parameters, tracks their 3D motion, and reconstructs static and moving geometry in canonical space, yielding simulation-compatible meshes. Across new real and simulated benchmarks, FunRec surpasses prior work by a large margin, achieving up to +50 mIoU improvement in part segmentation, 5-10 times lower articulation and pose errors, and significantly higher reconstruction accuracy. We further demonstrate applications on URDF/USD export for simulation, hand-guided affordance mapping and robot-scene interaction.