Skip to content

Dataset · BitRobot

Full-Body Motion Capture and Egocentric Video for Robots

BitRobot/robomocap

Multimodal dataset featuring optical motion capture and egocentric video for humanoid robot learning.

What's inside

Combines text, video, and 3D full-body optical motion capture data with first-person recordings. Filmed using six head-mounted and two wrist-mounted cameras, covering diverse locomotion and manipulation tasks. Dataset size ranges from 1,000 to 10,000 examples.

What you can do with it

  • Training control policies and locomotion for humanoid robots
  • Developing teleoperation systems and human-to-robot motion transfer
  • Training multimodal models to understand first-person video actions
Size
1K–10K rows
Downloads this month
61.4K
Likes
8

How to load it

from datasets import load_dataset

ds = load_dataset("BitRobot/robomocap")