Skip to content

Dataset · LightwheelAI

90k hours of egocentric video

LightwheelAI/EgoStandard

Egocentric videos with synchronized 3D hand poses for training and evaluating video models.

What's inside

EgoStandard is part of the EgoSuite-Open100K line featuring 90,000 hours of head-mounted video. It pairs egocentric video with synchronized 3D hand poses, while the body subset adds full-body poses. The data is in English, distributed via the LightwheelAI/EgoStandard bucket, with the Git repository serving as the card and access point.

What you can do with it

  • Train a video classifier or action recognition model
  • Evaluate 3D hand pose estimation on egocentric video
  • Fine-tune LoRA for video understanding tasks
Tasks
Video classification
Languages
English
Size
10K–100K rows
Downloads this month
791
Likes
379
Access
on request on Hugging Face

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("LightwheelAI/EgoStandard")