PhysForge: Generating Physics-Grounded 3D Assets for Interactive Virtual World
Summary
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.
View Cached Full Text
Cached at: 05/08/26, 08:11 AM
Paper page - PhysForge: Generating Physics-Grounded 3D Assets for Interactive Virtual World
Source: https://huggingface.co/papers/2605.05163
Abstract
PhysForge generates interactive 3D assets by combining visual-language modeling for physical planning with a physics-grounded diffusion model that synthesizes detailed geometry and kinematic parameters through a novel injection mechanism.
Synthesizing physics-grounded 3D assets is a critical bottleneck for interactive virtual worlds and embodied AI. Existing methods predominantly focus on static geometry, overlooking the functional properties essential for interaction. We propose that interactive asset generation must be rooted in functional logic and hierarchical physics. To bridge this gap, we introduce PhysForge, a decoupled two-stage framework supported by PhysDB, a large-scale dataset of 150,000 assets with four-tier physical annotations. First, a VLM acts as a “physical architect” to plan a “Hierarchical Physical Blueprint” defining material, functional, and kinematic constraints. Second, aphysics-grounded diffusion modelrealizes this blueprint by synthesizing high-fidelity geometry alongside precisekinematic parametersvia a novelKineVoxel Injection(KVI) mechanism. Experiments demonstrate that PhysForge produces functionally plausible,simulation-ready assets, providing a robust data engine for interactive 3D content and embodied agents.
View arXiv pageView PDFProject pageGitHub44Add to collection
Get this paper in your agent:
hf papers read 2605\.05163
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/2605.05163 in a model README.md to link it from this page.
Datasets citing this paper0
No dataset linking this paper
Cite arxiv.org/abs/2605.05163 in a dataset README.md to link it from this page.
Spaces citing this paper0
No Space linking this paper
Cite arxiv.org/abs/2605.05163 in a Space README.md to link it from this page.
Collections including this paper1
Similar Articles
PhysiFormer: Learning to Simulate Mechanics in World Space
PhysiFormer uses coordinate-space diffusion to generate physically-plausible 3D object motions without explicit inductive biases, enabling efficient multi-object reasoning and generalization to complex materials and geometries.
PhyGenHOI: Physically-Aware 4D Generation of Dynamic Human-Object Interactions
PhyGenHOI is a novel framework that generates physically accurate 4D human-object interactions by coupling motion diffusion models with material point method simulations using 3D Gaussian representations.
GamerForge
GamerForge is an AI-powered tool that transforms game, CGI, and VFX assets, allowing creators to enhance and edit digital assets efficiently.
@ziqi_huang_: An interesting work on Physical AI: PhysX-Omni. First unified sim-ready generation framework for rigid, deformable, and…
PhysX-Omni is a unified framework for simulation-ready physical 3D generation covering rigid, deformable, and articulated objects, with a new dataset (PhysXVerse) and benchmark (PhysX-Bench).
PhyMotion: Structured 3D Motion Reward for Physics-Grounded Human Video Generation
PhyMotion proposes a physics-grounded reward system that evaluates kinematic plausibility, contact consistency, and dynamic feasibility of human motion in generated videos, achieving stronger correlation with human judgment and improving motion realism in RL-based post-training.