Year: 2025
Type: preprint
Abstract: We introduce PAT3D, the first physics-augmented text-to-3D scene generation framework that integrates vision-language models with physics-based simulation to produce physically plausible, simulation-ready, and intersection-free 3D scenes. Given a text prompt, PAT3D generates 3D objects, infers their spatial relations, and organizes them into a hierarchical scene tree, which is then converted into initial conditions for simulation. A differentiable rigid-body simulator ensures realistic object interactions under gravity, driving the scene toward static equilibrium without interpenetrations. To further enhance scene quality, we introduce a simulation-in-the-loop optimization procedure that guarantees physical stability and non-intersection, while improving semantic consistency with the input prompt. Experiments demonstrate that PAT3D substantially outperforms prior approaches in physical plausibility, semantic consistency, and visual quality. Beyond high-quality generation, PAT3D uniquely enables simulation-ready 3D scenes for downstream tasks such as scene editing and robotic manipulation. Code and data are available at: https://github.com/Simulation-Intelligence/PAT3D.
Source: arXiv (Cornell University)
Authors: Lin, Guying, Huang, Kemeng, Liu, Michael, Gao, Ruihan, Chen, Hanke, Chen, Lyuhao, Lu, Beijia, Komura, Taku, Liu, Yuan, Zhu, Jun-Yan, Li, Minchen
Subfield: Computational Mechanics
Field: Engineering
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W4416975487
Cites:
Cited by:
Related to: