Data for the report "Mobile Application for Tunisian Sign Language (TunSL) Recognition".
Published: 14 January 2023 | Version 1 | DOI: 10.13140/RG.2.2.31806.36162
Contributors: Wassim Chakroun, Rihab Jerbi
Higher School of Communications of Tunis, SUP'COM
Image Processing, Image Classification, Sign Language, Tunisian Language, Recognition
This is a dataset of hand gestures that represent tunisian sign language. It is the first ever dataset of TunSL, it was collected in collaboration with Tunisian Association of Sign Language Interpreters (ATILS) as one of the tasks for the developement of a mobile app for the recognition of TunSL.
The dataset consists of 4423 images of TunSL common words performed by 7 people for 57 standard Tunisian signs. The number of images per class differs from one class to another with each image having the size of 224 × 224 pixels. These images were taken in different environments using laptop cameras with different backgrounds, clothes, hand angles, brightness level, hand accessories...
The folders and images have been named in Tunisian dialect.
The order of the dataset is as follows:
| Category | Signs | Images |
|---|---|---|
| Demandes | 24 | 1400 |
| Destinations | 8 | 608 |
| Famille | 12 | 1748 |
| Jours | 7 | 365 |
| Transport | 6 | 302 |
@{Wassim2022,
title = {(First ever) Tunisian Sign Language Dataset},
author = {Wassim Chakroun and Rihab Jerbi},
booktitle = {Mobile Application for Tunisian Sign Language (TunSL) Recognition},
address = {Tunis, Tunisia},
month = {April},
year = {2022}
}
[CC0-1.0](https://choosealicense.com/licenses/cc0-1.0/)