RGB-D Manipulation Dataset
Paired RGB-D recordings of robot manipulation with 3D grasp annotations and force measurements for training depth-aware grasping policies.
Dataset at a Glance
Comparison with Public Datasets
How Claru's dataset compares to publicly available alternatives.
| Dataset | Clips | Hours | Modalities | Environments | Annotations |
|---|---|---|---|---|---|
| GraspNet-1Billion | 97K grasps | ~10 | RGB-D | 190 scenes (lab) | 6-DoF grasps |
| ACRONYM | 8.8M grasps | N/A (sim) | Synthetic depth | Simulated | Grasp labels |
| Claru RGB-D Manipulation | 65K+ | 400+ | RGB-D, F/T | 30+ real setups | 3D grasps, poses, forces, trajectories |
Use Cases
6-DoF Grasp Planning
Predicting stable grasps using depth for geometry estimation. Example models: GraspNet, Contact-GraspNet, AnyGrasp.
Depth-Conditioned Policies
Policies using depth for spatial reasoning and collision avoidance. Example models: PerAct, 3D Diffusion Policy, RVT.
Object Pose Estimation
Estimating 6-DoF poses from RGB-D for manipulation planning. Example models: FoundationPose, MegaPose, BundleSDF.
Key References
- [1]Fang et al.. “GraspNet-1Billion: A Large-Scale Benchmark for General Object Grasping.” CVPR 2020, 2020. Link
- [2]Sundermeyer et al.. “Contact-GraspNet: Efficient 6-DoF Grasp Generation.” ICRA 2021, 2021. Link
- [3]Wen et al.. “FoundationPose: Unified 6D Pose Estimation and Tracking.” CVPR 2024, 2024. Link
How Claru Delivers This Data
Claru captures RGB-D manipulation data with diverse real objects — not just YCB benchmarks. Force/torque sensing adds the contact dynamics dimension pure depth datasets lack.
Frequently Asked Questions
Factory-calibrated RealSense cameras with sub-pixel alignment, validated per-session with checkerboard targets.
500+ unique objects spanning household items, tools, food, and industrial components in varying materials and sizes.
Force/torque contact detection combined with depth-based 3D localization, recording 6-DoF approach poses and grasp types.
Request a Sample Pack
Get a curated sample of rgb-d manipulation data with full annotations to evaluate for your project.