Skip to content

Dataset · zhicao

Egocentric manipulation videos with finger force sensors

zhicao/EgoForce

Synchronized video and tactile glove readings to transfer human fine motor skills to robots.

What's inside

A large dataset (10k to 100k records) presented as part of the work on the FACT multimodal model. Each record combines head-mounted stereo video and dual wrist camera videos with glove data—touch force, pressure, bending, and orientation of each finger on both hands—aligned frame-by-frame. Collected for training robotic systems.

What you can do with it

  • Training multimodal robot manipulator control models using tactile feedback
  • Classifying hand actions in egocentric video
  • Predicting touch force and finger positions from visual data
Tasks
Robotics, Video classification
Size
10K–100K rows
Downloads this month
13.6K
Likes
3

How to load it

from datasets import load_dataset

ds = load_dataset("zhicao/EgoForce")