This repository hosts a modular collection of pose-based datasets for action and sign language recognition. All datasets are preprocessed into .hdf5 format with 135 pose keypoints.
🧹 Structure
pose-action-recognition/
├── ISLR/ # Isolated Sign Language Recognition
│ ├── ASLCITIZEN/
│ │ ├── ASLCitizen100/
│ │ │ ├── ASLCitizen100_135-Train.hdf5
│ │ │ ├── ASLCitizen100_135-Val.hdf5
│ │ │… See the full description on the dataset page: https://huggingface.co/datasets/CristianLazoQuispe/pose-action-recognition.