| Issue |
MATEC Web Conf.
Volume 417, 2025
2025 RAPDASA-RobMech-PRASA-AMI Conference: Bridging the Gap between Industry & Academia - The 26th Annual International RAPDASA Conference, joined by RobMech, PRASA and AMI, co-hosted by CSIR and Tshwane University of Technology, Pretoria
|
|
|---|---|---|
| Article Number | 10004 | |
| Number of page(s) | 13 | |
| Section | Pattern Recognition | |
| DOI | https://doi.org/10.1051/matecconf/202541710004 | |
| Published online | 25 November 2025 | |
A real-time American Sign Language (ASL) alphabet recognition system for assistive communication using TinyML
School of Mechanical, Industrial and Aeronautical Engineering, University of the Witwatersrand, South Africa
* Corresponding author: Makhosazana.Moyo@wits.ac.za
Sign language is a critical communication tool for individuals with hearing impairments. This study focuses on developing a hand gesture recognition system to identify the American Sign Language (ASL) static alphabet using a Convolutional Neural Network (CNN) model deployed on a TinyML kit. The system is designed for real-time classification, making it a practical and accessible assistive technology. The model was tested at different distances, achieving an accuracy of 93.62% at 17.5 cm and 89.47% at 23 cm. Future work includes extending the system to recognise dynamic gestures and conducting user evaluations to improve robustness and accuracy.
© The Authors, published by EDP Sciences, 2025
This is an Open Access article distributed under the terms of the Creative Commons Attribution License 4.0, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.