Dataset · brownu
Multiview video and tactile data for robotics
brownu/deform360
Over 215 hours of recordings from 41 cameras and tactile grasps interacting with 198 deformable objects.
What's inside
A large-scale dataset from Brown University for studying deformable objects in robotics. It contains 1,980 interaction episodes with 198 everyday objects, captured by 41 surround-view cameras and dual-arm tactile manipulators (over 215 hours of video and text data).
What you can do with it
- Training world models to predict soft object deformation
- Training visual-tactile control algorithms for dual-arm robots
- Tasks
- Robotics
- Downloads this month
- 203.6K
- Likes
- 13
Models trained on it
How to load it
from datasets import load_dataset
ds = load_dataset("brownu/deform360")