melly777cc-by-4.0

DM-Tac Real Raw Recordings

Policy-allow-listed raw acquisition sessions from the DM-Tac 13-class real-data benchmark containing 166 sessions of tactile and RGB-D sensor data totaling approximately 103.5 GiB.

Downloads85
Episodes166

Why This Matters for Physical AI

This dataset provides high-fidelity tactile and RGB-D sensor recordings with calibrated ground-truth poses and object geometry, enabling research in multimodal perception and tactile-based robotic manipulation.

Technical Profile

Modalities
tactilergbdepth
Environment
lab
Task Types
tactile sensingobject classification
Episodes
166
Data Format
HDF5
Annotation Types
object_pose_basesensor_posescalibration_metadata
License
cc-by-4.0
Part of the DM-Tac family

Access

Need custom tactile data?

Claru builds purpose-built datasets for lab applications with dense human annotations and quality assurance.

Request a Sample Pack

Related Datasets