ASLA-Leap dataset is an American Sign Language Alphabet dataset collected with a Leap Motion Controller. It has 26 signs performed by 5 subjects and 500 samples for each sign.
Download (8.9GB .zip)
This notebook will show you how to read in the recorded Leap data.
Also, you can find the dataset in .jpg
and .csv
files in the /dataset
folder, which is more convenient to get started.
Please cite our paper in your publications if it helps your research.
@article{tao2018asl,
title = {American Sign Language Alphabet Recognition Using Leap Motion Controller},
journal = {Proceedings of the 2018 IISE Annual Conference},
pages = {599-604},
author = {Wenjin Tao and Ze-Hao Lai and Ming C. Leu and Zhaozheng Yin},
note = {IISE Annual Conference and Expo 2018},
year = {2018}
}
If you are interested in this topic, here's another study we did for American Sign Language Alphabet Recognition.
@article{tao2018american,
title={American Sign Language alphabet recognition using Convolutional Neural Networks with multiview augmentation and inference fusion},
author={Tao, Wenjin and Leu, Ming C and Yin, Zhaozheng},
journal={Engineering Applications of Artificial Intelligence},
volume={76},
pages={202--213},
year={2018},
publisher={Elsevier}
}