Show HN: Impressive hand detection AI algorithm

Hand-action recognition on any video with Hiera-Hand (ChildPlay-Hand, ECCVW 2024). It tracks people, localizes their hands, and labels every hand in every frame as grasp, hold, operate, or release.

idiap/childplay_hand.

⚠️ This is a standalone demo based on the original research work; for training, evaluation, and the dataset, see

./setup_env.sh              # needs uv; Python 3.11 venv in .venv
source .venv/bin/activate

Fetch the checkpoints from Zenodo into checkpoints/:

python src/download_checkpoints.py              # manipulation (~0.4 GB download)
python src/download_checkpoints.py --task all   # + object (~0.8 GB download)
python src/demo.py input.mp4 --output result.mp4

Works best on clips where people are fully visible, without camera cuts.

1. Track people and their pose : YOLO26m-pose + BoT-SORT, in one pass.
2. Find hands : each hand box sits just past the wrist, along the elbow→wrist direction.
3. Recognize : for every hand and frame, a ~1 s window (32 frames, 16 sampled) is cropped around the hand at 224×224 and fed toHiera-Base (51M params, 205 MB; MAE-pretrained on Kinetics-400, fine-tuned on ChildPlay-Hand), which outputs background / grasp / hold / operate / release.
4. Display : smoothed hand boxes and hand actions.

Note: The paper used HRNet-W32 for pose; this demo uses YOLO26m-pose for speed, so predictions may differ slightly from the reported results.

• Code : GPL-3.0, based onidiap/childplay_hand (© Idiap Research Institute).
• Checkpoints : CC BY-NC 4.0 (non-commercial), fromZenodo .
• Pose model :Ultralytics YOLO26, AGPL-3.0.
• Hiera architecture (src/hiera/ ): Apache-2.0, © Meta.

@inproceedings{Farkhondeh_ECCVW_2024,
  author    = {Farkhondeh*, Arya and Tafasca*, Samy and Odobez, Jean-Marc},
  title     = {ChildPlay-Hand: A Dataset of Hand Manipulations in the Wild},
  booktitle = {Proceedings of the European Conference on Computer Vision (ECCV) Workshops},
  year      = {2024},
  note      = {* Equal contribution}
}