Datasets:

Modalities:
Video
Size:
< 1K
ArXiv:
Tags:
video
Libraries:
Datasets
License:
Dataset Viewer
Auto-converted to Parquet Duplicate
Search is not available for this dataset
video
video
8
62
label
class label
8 classes
0dance
0dance
1egoexo
1egoexo
1egoexo
1egoexo
2fab4
2fab4
2fab4
3humanoid
3humanoid
3humanoid
4retarget
4retarget
4retarget
4retarget
4retarget
4retarget
4retarget
4retarget
5scenes
6teaser
7world
7world
7world
7world
7world
7world
7world

FAB project page media

The videos played on the FAB project page, www.jinkuncao.com/FAB. Code: github.com/noahcao/FAB · Paper: arXiv 2611.25801.

These are demo videos only, not training or evaluation data. Each clip keeps the licence of its source; overlays and renders are ours:

  • EMDB, RICH, EgoBody frames: under each dataset's licence, for non-commercial research. Bystander faces blurred.
  • CORE4D-Real: CC BY 4.0. Participants' faces blurred by CORE4D, other heads blurred by us.
  • Pexels clips: Pexels License. Clip credits are listed on the project page.
  • Wikimedia Commons photos: CC BY 2.0. HoloAssist: CDLA-Permissive-2.0. Egocentric-10K: Apache-2.0.
  • Unitree G1 model: BSD 3-Clause (unitree_ros).

The full credits are in the Credits section of the project page.

@misc{fab2026,
  title         = {Fast and Accurate Body Estimation in Pixels, World Space and Scenes},
  author        = {Cao, Jinkun},
  year          = {2026},
  eprint        = {2611.25801},
  archivePrefix = {arXiv},
  primaryClass  = {cs.CV},
  url           = {https://arxiv.org/abs/2611.25801}
}
Downloads last month
115

Paper for noahcao/FAB-page