Research map / Video / Video generation
PhysGen: Rigid-Body Physics-Grounded Image-to-Video Generation
Shaowei Liu, Zhongzheng Ren, Saurabh Gupta, Shenlong Wang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| estimate physical parameters with GPT-4V instead of random values from human-estimated ranges (ablation) | PhysGen · collected GT scene videos | Image-FID: 111.01 -> 105.70 | Improved |
| estimate physical parameters with GPT-4V instead of random values from human-estimated ranges (ablation) | PhysGen · collected GT scene videos | Motion-FID: 36.60 -> 30.20 | Improved |
| ground object dynamics in rigid-body simulation instead of purely data-driven I2V | PhysGen vs SEINE, DynamiCrafter, I2VGen-XL · human evaluation | physical-realism (1-5): 2.11 -> 4.14 | Improved |
| ground dynamics in physics simulation versus best baseline | PhysGen · collected GT scene videos | Motion-FID: 38.76 -> 30.20 | Improved |
| physics-grounded generation versus DynamiCrafter on appearance | PhysGen · collected GT scene videos | Image-FID: 99.64 -> 105.70 | Hurt |
| add generative diffusion refinement after composition and relighting (ablation) | PhysGen rendering · qualitative | photo-realism | Improved |
| add relighting to composited frames (ablation) | PhysGen rendering · qualitative | lighting realism | Mixed |
| diffusion refinement injects generated latent into foreground | PhysGen · qualitative limitation | foreground fidelity | Hurt |
Later papers that compared with this method
2 later accepted papers report PhysGen as a baseline in their results tables; 2 of them report a better number on most of the results they share with it.
- NS-Diff: Fluid Navier-Stokes Guided Video Diffusion via Reinforcement Learning (CVPR 2026 · better on 4 of 4 shared results)
- WonderPlay: Dynamic 3D Scene Generation from a Single Image and Actions (ICCV 2025 · better on 5 of 5 shared results)
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- SEINE — SEINE: Short-to-Long Video Diffusion Model for Generative Transition and Prediction (ICLR 2024) · better on 4 of 4 shared results
- DynamiCrafter — DynamiCrafter: Animating Open-domain Images with Video Diffusion Priors (ECCV 2024) · better on 3 of 4 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- PhysCtrl: Generative Physics for Controllable and Physics-Grounded Video Generation (NeurIPS 2025 · similarity 0.93)
- PhysChoreo: Physics-Controllable Video Generation with Part-Aware Semantic Grounding (ECCV 2026 · similarity 0.92)
- PhysGen3D: Crafting a Miniature Interactive World from a Single Image (CVPR 2025 · similarity 0.89)
- Force Prompting: Video Generation Models Can Learn And Generalize Physics-based Control Signals (NeurIPS 2025 · similarity 0.88)
- PhysAnimator: Physics-Guided Generative Cartoon Animation (CVPR 2025 · similarity 0.88)
- How Far Is Video Generation from World Model: A Physical Law Perspective (ICML 2025 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
