Skip to content

Dataset · Exploration-Lab

Indian Sign Language processing benchmark

Exploration-Lab/iSign

Dataset of over 100,000 records for Indian Sign Language recognition and translation.

What's inside

Contains data for five Indian Sign Language processing tasks: video-to-text and pose translation, pose generation from text, word prediction, and gesture semantics determination. The dataset includes videos, .pose files, and text translations in CSV. Size ranges from 100,000 to 1 million examples, with data open for non-commercial research only.

What you can do with it

  • Training sign language to text and reverse translation models (SignVideo2Text, Text2Pose)
  • Testing word recognition and semantic algorithms using skeletal poses
Tasks
Translation
Size
100K–1M rows
Downloads this month
2.2K
Likes
54
Access
on request on Hugging Face

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("Exploration-Lab/iSign")