MonoArt: Progressive Structural Reasoning for Monocular Articulated 3D Reconstruction

TL;DR

MonoArt introduces a progressive reasoning framework for monocular 3D reconstruction of articulated objects, achieving stable, accurate, and efficient inference without multi-stage pipelines or external templates.

Contribution

It proposes a unified, progressive structural reasoning approach that transforms visual data into canonical geometry and part representations, improving stability and interpretability.

Findings

Achieves state-of-the-art accuracy on PartNet-Mobility

Demonstrates fast inference speed

Generalizes to robotic manipulation and scene reconstruction

Abstract

Reconstructing articulated 3D objects from a single image requires jointly inferring object geometry, part structure, and motion parameters from limited visual evidence. A key difficulty lies in the entanglement between motion cues and object structure, which makes direct articulation regression unstable. Existing methods address this challenge through multi-view supervision, retrieval-based assembly, or auxiliary video generation, often sacrificing scalability or efficiency. We present MonoArt, a unified framework grounded in progressive structural reasoning. Rather than predicting articulation directly from image features, MonoArt progressively transforms visual observations into canonical geometry, structured part representations, and motion-aware embeddings within a single architecture. This structured reasoning process enables stable and interpretable articulation inference without external motion templates or multi-stage pipelines. Extensive experiments on PartNet-Mobility demonstrate that OM achieves state-of-the-art performance in both reconstruction accuracy and inference speed. The framework further generalizes to robotic manipulation and articulated scene reconstruction.