SNAP3D: Physically Grounded 3D Parts for Assembly from a Single Image
Summary
SNAP3D introduces a physics-guided framework to improve part-aware 3D generation from a single image, ensuring stable and physically valid assemblies through simulation feedback and resolving issues like inter-part penetration.
View Cached Full Text
Cached at: 09/14/26, 02:33 AM
Paper page - SNAP3D: Physically Grounded 3D Parts for Assembly from a Single Image
Source: https://huggingface.co/papers/2609.13146
Abstract
A physics-guided framework improves part-aware 3D generation by resolving inter-part penetration, recovering contact graphs, and refining parameterized connectors via simulation feedback to ensure stable, physically valid assemblies.
Part-aware 3D asset generation enables applications such as editing, articulation, simulation, and fabrication, yet existing methods can generate visually complete individual parts without ensuring that they form a valid physical assembly. Consequently, generated neighboring parts may interpenetrate, lack valid connections, or collapse under gravity. We propose a physics-guided framework for improving single-imagepart-aware 3D generationwith physically compatible geometry and stable connections. Our method resolvesinter-part penetration, recovers acontact graphbetween neighboring parts, and introducesparameterized connectorsat their contact surfaces. Using feedback fromphysical simulation, we refine connector placement, orientation, and dimensions to improveassembly stabilitywhile preserving the generated geometry. We further introduce aphysics-based evaluationprotocol that complements conventional geometric metrics by directly testing assembly validity and stability under gravity. Experiments comparing against multiple part-aware 3D generators show substantial improvements in physical realizability and stability while maintaining geometric quality. We additionally validate the resulting parts through 3D printing and real-world assembly.
View arXiv pageView PDFProject pageGitHub1Add to collection
Get this paper in your agent:
hf papers read 2609\.13146
Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash
Models citing this paper0
No model linking this paper
Cite arxiv.org/abs/2609.13146 in a model README.md to link it from this page.
Datasets citing this paper0
No dataset linking this paper
Cite arxiv.org/abs/2609.13146 in a dataset README.md to link it from this page.
Spaces citing this paper0
No Space linking this paper
Cite arxiv.org/abs/2609.13146 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
PhysForge: Generating Physics-Grounded 3D Assets for Interactive Virtual World
PhysForge is a two-stage framework that generates interactive 3D assets with grounded physics and kinematic parameters, addressing the bottleneck of static geometry in virtual worlds.
Sat3DGen: Comprehensive Street-Level 3D Scene Generation from Single Satellite Image
Sat3DGen introduces a geometry-first approach for generating street-level 3D scenes from a single satellite image, achieving improved geometric accuracy and photorealism through novel constraints and training strategies. The method demonstrates significant improvements over prior work on the VIGOR-OOD benchmark.
SNAP-FM: Sparse Nonlinear Accelerated Projection for Physics-Constrained Generative Modeling
Proposes SNAP-FM, a method that leverages sparse GPU nonlinear optimization to accelerate constraint projection in physics-constrained generative modeling, achieving faster inference while preserving exact physical constraint satisfaction.
Design 3D-printable parts by talking
nurb enables users to design 3D-printable parts through natural language conversation with an AI, handling design, checks, and revisions without traditional CAD software.
StampFormer: A Physics-Guided Material-Geometry-Coupled Multimodal Model for Rapid Prediction of Physical Fields in Sheet Metal Stamping
StampFormer is a physics-guided deep learning framework that fuses geometry and material properties to predict FEA outcomes for sheet metal stamping in under a second, achieving high fidelity with less than 8.5% relative error.