Artic-O: End-to-End Articulated Object Reconstruction via Latent Geometry Learning

2026-06-20

AI Summary

This work addresses the challenge of reconstructing articulated objects from sparse images, which requires jointly recovering geometric completeness, part segmentation, and motion parameters—a task often hindered by poor consistency and inefficiency in existing multi-stage approaches. The authors propose Artic-O, an end-to-end feedforward framework that, for the first time, enables articulated object reconstruction using implicit geometric representations. By leveraging a frozen flow-matching decoder, Artic-O maps multi-state sparse observations into a pretrained geometric space, fusing visual tokens, geometric latent variables, and point-wise features to simultaneously complete shape, segment parts, and predict joint parameters. A geometry-articulation curriculum learning strategy and a decoupled dual-path training scheme are introduced to enhance geometric-motion consistency and inference efficiency. On PartNet-Mobility, Artic-O achieves lower Chamfer Distance, higher F-score, comparable or superior joint accuracy, and reduces per-object inference time from 9 minutes to 0.3 seconds.

Abstract

Reconstructing articulated objects from sparse images requires recovering complete geometry, movable parts, and motion parameters. Recent methods typically separate geometry reconstruction, part reasoning, and articulation estimation into different stages. This separation can weaken consistency between shape, active parts, and motion, while also incurring substantial inference cost. We introduce Artic-O, an end-to-end, feed-forward framework for articulated object reconstruction via latent geometry learning. Instead of fitting geometry in image or view space, Artic-O maps sparse multi-state observations into a pretrained latent geometry space, where a frozen flow-matching decoder provides a complete-shape prior for recovering visible and occluded structures. To connect geometry with articulation, Artic-O fuses visual tokens, geometry latents, and point-wise decoder features in an image-grounded part-reasoning module for active-part segmentation and articulation prediction. We further train the model with a geometry-to-articulation curriculum and a decoupled two-pass strategy to balance reconstruction and part-level supervision. On PartNet-Mobility, Artic-O achieves strong reconstruction quality while being substantially more efficient than LARM, a strong prior method. It reduces Chamfer Distance, improves F-score, and achieves comparable or better articulation accuracy across most joint metrics, while reducing inference time from 9 minutes to about 0.3 seconds per object.

Problem

• articulated object reconstruction
• sparse images
• geometry recovery
• motion parameters
• part reasoning

Innovation

• latent geometry learning
• end-to-end articulated reconstruction
• flow-matching decoder
• image-grounded part reasoning
• geometry-to-articulation curriculum