MOVAS-3D — High-Fidelity 3D Assets & Environments
Reconstructed object meshes and full environment scans, physically scaled and simulation-ready for Isaac Sim, MuJoCo and Genesis.
Overview
Assets are reconstructed from the same physical objects that appear in the video corpora, which means a policy trained in simulation and a policy trained on real video are looking at the same object — not a visually similar stand-in.
Every mesh ships with a decomposed collision hull and measured physical properties, so it drops into a physics engine without manual authoring.
Data Structure
Directory layout as delivered. Every clip is self-contained: no cross-referencing required to train on a single sample.
movas-3d/
├── objects/{object_id}/
│ ├── mesh.usd
│ ├── mesh.glb
│ ├── collision.obj # convex decomposition
│ ├── material.json
│ └── physics.json # mass, CoM, inertia, friction
├── environments/{env_id}/
│ ├── scene.usd
│ └── navmesh.obj
└── manifest.parquetAnnotation Schema
Category
Open-vocabulary object class and material family.
Physics
Measured mass, centre of mass, inertia tensor, surface friction.
Affordance
Grasp candidate regions and articulation joints.
Quality Gates
Published thresholds, enforced at ingest. Records that fail are rejected or flagged in the manifest — never silently included.
| Gate | Threshold | Method |
|---|---|---|
| Dimensional accuracy | ± 2 mm | Verified against physical calipers on a 5% sample. |
| Watertight rate | > 99% | Meshes pass manifold check before release. |
| Sim import success | 100% | Every asset load-tested in Isaac Sim and MuJoCo. |
Formats & Access
Export formats
Modalities included
Cloud delivery
Direct to your S3, GCS or OSS bucket. Manifest-driven incremental sync.
Air-gapped
Physical media transfer for regulated or offline environments.
Via Movas-OS
Stream and re-annotate in place through the Movas-OS platform.
Use Cases
Sim-to-real training
Domain randomisation over assets that match the real capture set.
Digital twins
Environment scans for facility-level simulation.
Grasp synthesis
Affordance-annotated geometry for analytic and learned grasping.