GEAR: GEometry-motion Alternating Refinement for Articulated Object Modeling with Gaussian Splatting

2026-04-08

AI Summary

This work addresses the challenging problem of high-fidelity reconstruction of complex articulated objects, where strong coupling between geometry and motion hinders existing methods, leading to unstable joint optimization and limited generalization. To overcome this, we propose an EM-style alternating optimization framework within the Gaussian Splatting representation, modeling geometry and motion as interdependent components. By iteratively optimizing implicit part segmentation (E-step) and explicit joint parameters (M-step), our approach effectively decouples their joint learning. The method integrates multi-view 2D part segmentation priors with weakly supervised regularization, significantly improving reconstruction quality while preserving generalization. Evaluated on multiple benchmarks and our newly introduced GEAR-Multi dataset, our approach achieves state-of-the-art performance in both geometric reconstruction and motion estimation, particularly excelling on complex articulated objects with multiple moving parts.

Abstract

High-fidelity interactive digital assets are essential for embodied intelligence and robotic interaction, yet articulated objects remain challenging to reconstruct due to their complex structures and coupled geometry-motion relationships. Existing methods suffer from instability in geometry-motion joint optimization, while their generalization remains limited on complex multi-joint or out-of-distribution objects. To address these challenges, we propose GEAR, an EM-style alternating optimization framework that jointly models geometry and motion as interdependent components within a Gaussian Splatting representation. GEAR treats part segmentation as a latent variable and joint motion parameters as explicit variables, alternately refining them for improved convergence and geometric-motion consistency. To enhance part segmentation quality without sacrificing generalization, we leverage a vanilla 2D segmentation model to provide multi-view part priors, and employ a weakly supervised constraint to regularize the latent variable. Experiments on multiple benchmarks and our newly constructed dataset GEAR-Multi demonstrate that GEAR achieves state-of-the-art results in geometric reconstruction and motion parameters estimation, particularly on complex articulated objects with multiple movable parts.

Problem

• articulated object modeling
• geometry-motion coupling
• 3D reconstruction
• Gaussian Splatting
• part segmentation

Innovation

• Gaussian Splatting
• articulated object modeling
• alternating optimization
• geometry-motion consistency
• weakly supervised segmentation