SimuScene: Simulation-Ready Compositional 3D Scene Reconstruction from a Single Image
SimuScene: Simulation-Ready Compositional 3D Scene Reconstruction from a Single Image
This paper presents SimuScene, a method for reconstructing interactive, simulation-ready 3D scenes from a single image by integrating physics into the shape and layout estimation process. Unlike post-hoc corrections, SimuScene uses a physics engine as a diagnostic tool during generation to convert penetration and support failures into quantitative signals that drive amodal shape resampling and gravity-axis stretching. Extensive experiments show state-of-the-art performance on physical stability and geometric alignment benchmarks, with demonstrated utility in humanoid control and robot-arm manipulation tasks.
SimuScene introduces a physics-informed generative pipeline that uses gravity-based diagnostic simulation to correct geometric errors during 3D scene reconstruction, achieving improved physical stability and geometric alignment in benchmarks while enabling deployment in humanoid control and robot-arm manipulation tasks.
Source: arXiv Robotics — research abstracts · Read original article ↗
Loading article text…
What the source reports
Publisher-reported claims, with original evidence. These results have not been independently verified by RoboSignal.
- Environment
- SimulationOpen source S2
- paper: not_reported
- code: not_reported
- weights: not_reported
- dataset: not_reported
Source excerpts and review record
Automatically extracted; no manual editorial approval recorded.
arXiv:2606.03994v2 Announce Type: replace-cross Abstract: Reconstructing interactive, simulation-ready 3D scenes from a single image is a critical bottleneck for robotic manipulation. While recent single-image lifters recover plausible per-object shapes, composing them yields scenes that collapse under physical simulation due to interpenetrating, hovering, or sinking objects. Existing physics-aware compositional reco
Open source S2
Source：arXiv Robotics — research abstracts · arxiv.org