GigaHands: A Massive Annotated Dataset of Bimanual Hand Activities

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Fu, Rao, Zhang, Dingxi, Jiang, Alex, Fu, Wanjia, Funk, Austin, Ritchie, Daniel, Sridhar, Srinath
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909572463067136
author Fu, Rao
Zhang, Dingxi
Jiang, Alex
Fu, Wanjia
Funk, Austin
Ritchie, Daniel
Sridhar, Srinath
author_facet Fu, Rao
Zhang, Dingxi
Jiang, Alex
Fu, Wanjia
Funk, Austin
Ritchie, Daniel
Sridhar, Srinath
contents Understanding bimanual human hand activities is a critical problem in AI and robotics. We cannot build large models of bimanual activities because existing datasets lack the scale, coverage of diverse hand activities, and detailed annotations. We introduce GigaHands, a massive annotated dataset capturing 34 hours of bimanual hand activities from 56 subjects and 417 objects, totaling 14k motion clips derived from 183 million frames paired with 84k text annotations. Our markerless capture setup and data acquisition protocol enable fully automatic 3D hand and object estimation while minimizing the effort required for text annotation. The scale and diversity of GigaHands enable broad applications, including text-driven action synthesis, hand motion captioning, and dynamic radiance field reconstruction. Our website are avaliable at https://ivl.cs.brown.edu/research/gigahands.html .
format Preprint
id arxiv_https___arxiv_org_abs_2412_04244
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle GigaHands: A Massive Annotated Dataset of Bimanual Hand Activities
Fu, Rao
Zhang, Dingxi
Jiang, Alex
Fu, Wanjia
Funk, Austin
Ritchie, Daniel
Sridhar, Srinath
Computer Vision and Pattern Recognition
Understanding bimanual human hand activities is a critical problem in AI and robotics. We cannot build large models of bimanual activities because existing datasets lack the scale, coverage of diverse hand activities, and detailed annotations. We introduce GigaHands, a massive annotated dataset capturing 34 hours of bimanual hand activities from 56 subjects and 417 objects, totaling 14k motion clips derived from 183 million frames paired with 84k text annotations. Our markerless capture setup and data acquisition protocol enable fully automatic 3D hand and object estimation while minimizing the effort required for text annotation. The scale and diversity of GigaHands enable broad applications, including text-driven action synthesis, hand motion captioning, and dynamic radiance field reconstruction. Our website are avaliable at https://ivl.cs.brown.edu/research/gigahands.html .
title GigaHands: A Massive Annotated Dataset of Bimanual Hand Activities
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2412.04244