Part$^{2}$GS: Part-aware Modeling of Articulated Objects using 3D Gaussian Splatting

2025-06-20

AI Summary

To address low geometric fidelity and physically inconsistent motion in digital twins of multi-component articulated objects, this paper proposes a part-aware 3D Gaussian Splatting representation framework. Our method is the first to jointly integrate part-wise decoupled modeling with physics-guided motion optimization: it introduces an explicit repulsive point field to model inter-part collision constraints; designs three physical priors—contact preservation, velocity consistency, and vector field alignment; and enables structured motion learning via differentiable joint parameter optimization. Evaluated on both synthetic and real-world datasets, our approach achieves up to a 10× reduction in part-level Chamfer distance over state-of-the-art methods. It yields more stable motion trajectories and preserves finer geometric details, significantly enhancing the structural plausibility and physical credibility of digital twin models.

Technology Category

• Computer Vision: Low Level & Physics-based Vision
• Intelligent Robots: Motion and Path Planning
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• User Modeling, Personalization and Recommendation: User modeling and simulation for interactive and conversational systems

Abstract

Articulated objects are common in the real world, yet modeling their structure and motion remains a challenging task for 3D reconstruction methods. In this work, we introduce Part$^{2}$GS, a novel framework for modeling articulated digital twins of multi-part objects with high-fidelity geometry and physically consistent articulation. Part$^{2}$GS leverages a part-aware 3D Gaussian representation that encodes articulated components with learnable attributes, enabling structured, disentangled transformations that preserve high-fidelity geometry. To ensure physically consistent motion, we propose a motion-aware canonical representation guided by physics-based constraints, including contact enforcement, velocity consistency, and vector-field alignment. Furthermore, we introduce a field of repel points to prevent part collisions and maintain stable articulation paths, significantly improving motion coherence over baselines. Extensive evaluations on both synthetic and real-world datasets show that Part$^{2}$GS consistently outperforms state-of-the-art methods by up to 10$ imes$ in Chamfer Distance for movable parts.

Problem

• Modeling articulated objects with high-fidelity geometry
• Ensuring physically consistent motion in 3D reconstruction
• Preventing part collisions in articulated object modeling

Innovation

• Part-aware 3D Gaussian representation for articulated objects
• Physics-based constraints for motion consistency
• Field of repel points to prevent part collisions