DROID - Evaluation on objects that are unseen in both simulation and real world
Explore the reconstructed 3D scenes interactively. Drag to rotate, scroll to zoom.
Note: 3D viewers are not supported on mobile. Please visit on desktop for the interactive experience.
SimFoundry scenes can be used to evaluate real-world policies in simulation, with a mean Pearson correlation of 0.911.
Side-by-side comparison of policy execution in real world versus SimFoundry.
Policy: DreamZero
Policy: Gr00t N1.7
Policy: π0.5
Policy: π0.5
Policy: π0.5
SimFoundry predictions closely track real-world performance and outperform a State-of-the-Art baseline (PolaRiS, Jain et al. 2025)
(hover over the points to see additional metrics)
Higher Pearson r and lower MMRV indicate better sim-to-real correlation
Policies Trained on SimFoundry Data Transfer Zero-Shot to Real-World.
Evaluation results on the DROID platform in the real world
SimFoundry Digital Cousins Enable Policy Generalization to Novel Objects and Tasks
Selected results across diverse scenes and tasks.
SimFoundry data boosts real-world VLA performance
SimFoundry outperforms State-of-the-Art methods on scene reconstruction accuracy
Click an object in the real scene to view its reconstructed, physics-ready mesh.
SimFoundry recovers more accurate object meshes and poses than SAM3D, especially on occluded, cluttered scenes. The pipeline also generalizes across input modalities, supporting open-source datasets and synthetically generated images out of the box.
Note: 3D viewers are not supported on mobile. Please visit on desktop for the interactive experience.
SimFoundry erases foreground objects to produce a background-only video, which is refined and used to train a 3D Gaussian Splat for high-fidelity background reconstruction.
SimFoundry extracts per-object relevant information (segmentation masks, depth, etc.), generates 3D visual meshes via 2D-to-3D generation models, and compiles the final output scene by annotating relevant physical parameters and sanity checking the overall scene configuration in a physics simulator. SimFoundry additionally supports diverse simulated augmentations of objects, scenes, and tasks. SimFoundry's modular design ensures that as individual foundation models improve, the pipeline improves with them—requiring no redesign, only a component swap.
