WonderPlay: Dynamic 3D Scene Generation from a Single Image and Actions
Li, Zizhang, Yu, Hong-Xing, Liu, Wei, Yang, Yin, Herrmann, Charles, Wetzstein, Gordon, Wu, Jiajun
–arXiv.org Artificial Intelligence
WonderPlay is a novel framework integrating physics simulation with video generation for generating action-conditioned dynamic 3D scenes from a single image. While prior works are restricted to rigid body or simple elastic dynamics, WonderPlay features a hybrid generative simulator to synthesize a wide range of 3D dynamics. The hybrid generative simulator first uses a physics solver to simulate coarse 3D dynamics, which subsequently conditions a video generator to produce a video with finer, more realistic motion. The generated video is then used to update the simulated dynamic 3D scene, closing the loop between the physics solver and the video generator. This approach enables intuitive user control to be combined with the accurate dynamics of physics-based simulators and the expressivity of diffusion-based video generators. Experimental results demonstrate that WonderPlay enables users to interact with various scenes of diverse content, including cloth, sand, snow, liquid, smoke, elastic, and rigid bodies -- all using a single image input. Code will be made public. Project website: https://kyleleey.github.io/WonderPlay/
arXiv.org Artificial Intelligence
Dec-2-2025
- Country:
- Asia > Middle East
- Saudi Arabia > Northern Borders Province > Arar (0.04)
- North America > United States
- Utah (0.04)
- Asia > Middle East
- Genre:
- Research Report (0.70)
- Industry:
- Leisure & Entertainment (0.46)
- Technology:
- Information Technology > Artificial Intelligence
- Machine Learning (1.00)
- Natural Language (0.68)
- Representation & Reasoning (0.89)
- Vision (1.00)
- Information Technology > Artificial Intelligence