jaibrdhn2026cc-by-4.0
DROID-3D: Optimized Camera Calibration for DROID
Recalibrated multi-view camera extrinsics and per-episode intrinsics for the DROID robot-manipulation dataset, with dense metric depth recovered from stereo recordings. The calibration uses joint factor graph optimization across 71,006 episodes to recover shared kinematic parameters and per-scene extrinsics.
Downloads22
Episodes71,006
Why This Matters for Physical AI
This dataset provides precisely calibrated 3D camera geometry and metric depth for the large-scale DROID manipulation dataset, enabling accurate 3D perception and spatial reasoning for robotic learning systems.
Technical Profile
- Modalities
- rgbdepthproprioception
- Robot Embodiments
- Franka Panda
- Environment
- lab
- Task Types
- manipulation
- Episodes
- 71,006
- Data Format
- JSONL
- Annotation Types
- camera_calibrationextrinsicsintrinsicspseudo_labels
- License
- cc-by-4.0
Access
Need custom rgb data?
Claru builds purpose-built datasets for lab applications with dense human annotations and quality assurance.
Request a Sample Pack