Articulated Object Reconstruction from Rest-State Observation

Hugging Face Daily Papers Papers

Summary

This paper introduces a rest-state framework that reconstructs articulated objects from a single closed configuration, using explicit meshes, vision-language outputs, and video diffusion models to generate and validate articulation hypotheses without observed motion.

Building interactive digital twins requires recovering both 3D geometry and the kinematic structures that govern how objects articulate. Yet existing methods for articulated object reconstruction require explicitly observable motion from multiple articulation states. We introduce a rest-state formulation that reconstructs articulated objects from a single closed configuration, an inherently ill-posed setting where geometry, semantics, and motion priors compensate for the absence of motion cues. Our framework adopts an explicit mesh as an intermediate representation for cross-model verification and fusion, reconciling noisy outputs from vision-language and segmentation models into spatially consistent part structures. To estimate joint parameters without observed motion, we use a video diffusion model to synthesize articulation hypotheses and validate them through geometric consistency. Our approach achieves accurate part decomposition and physically plausible articulation, performing competitively with motion-observing reconstruction-based, generation-based, and modular pretrained-model baselines.
Original Article
View Cached Full Text

Cached at: 08/12/26, 08:19 AM

Paper page - Articulated Object Reconstruction from Rest-State Observation

Source: https://huggingface.co/papers/2607.27749

Abstract

A rest-state framework reconstructs articulated objects from a single closed configuration by fusing vision-language outputs into consistent part meshes and validating synthesized motion hypotheses via geometric consistency.

Building interactivedigital twinsrequires recovering both 3D geometry and the kinematic structures that govern how objects articulate. Yet existing methods forarticulated object reconstructionrequire explicitly observable motion from multiple articulation states. We introduce arest-state formulationthat reconstructs articulated objects from a single closed configuration, an inherently ill-posed setting where geometry, semantics, and motion priors compensate for the absence of motion cues. Our framework adopts anexplicit meshas an intermediate representation for cross-model verification and fusion, reconciling noisy outputs from vision-language andsegmentation modelsinto spatially consistent part structures. To estimatejoint parameterswithout observed motion, we use avideo diffusion modelto synthesizearticulation hypothesesand validate them throughgeometric consistency. Our approach achieves accurate part decomposition and physically plausible articulation, performing competitively with motion-observing reconstruction-based, generation-based, and modular pretrained-model baselines.

View arXiv pageView PDFProject pageAdd to collection

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2607.27749 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2607.27749 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2607.27749 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.

Similar Articles

Revisiting Articulated Parts Perception in Robot Manipulation

Hugging Face Daily Papers

This paper introduces Geometric Primary Structure (GPS), a new representation for articulated parts perception in robot manipulation, enabling efficient VR-based annotation and achieving a 73% success rate without fine-tuning.

@artemZholus: thanks! in the second paper (https://arxiv.org/abs/2605.06388) we used your (and RAE's) recipe and it worked.

X AI KOLs Following

This paper systematically compares reconstruction-based and semantic latent spaces for action-conditioned latent diffusion world models in robotics. It finds that semantic encoders like V-JEPA 2.1 generally outperform reconstruction encoders on policy-relevant metrics, advocating for semantic latent spaces as a stronger foundation for robotics world models.