The dataset contains 500 unique ISL (Indian Sign Language) words recorded across 15 users, along with pose and mediapipe landmark extractions for each session.
Dataset Structure
ISL-DATA/
├── Videos/
│ ├── User001/
│ │ └── .mp4
│ ├── User002/
│ │ ...
│ └── User015/
├── Landmarks/
│ ├── MediaPipe/
│ │ ├── User001/
│ │ │… See the full description on the dataset page: https://huggingface.co/datasets/ISL500/ISL-DATA.
