Location
Kumasi, Ghana & Virtual
Start Date
6-8-2026 12:00 PM
End Date
6-8-2026 12:30 PM
Description
Deaf and Hard-of-Hearing (DHH) individuals often face communication challenges with people who do not understand sign language. Existing systems lack support for South African Sign Language (SASL) and are limited in accessibility, efficiency and accuracy. This paper introduces Ubuntu Signs, a mobile prototype tool for real-time SASL translation using a camera-based interface and machine learning. The tool uses two models to recognize gestures: a static model for alphabetical and numerical gestures and a dynamic model for continuous gestures representing words and phrases. A preliminary user study evaluated the effectiveness, efficiency, accuracy and satisfaction of Ubuntu Signs. Results indicated high usability and positive user satisfaction, particularly for the tool’s simplicity, intuitive design and accurate recognition of static gestures. However, limitations were observed in dynamic gesture recognition. Future work could focus on improving this model. This study contributes to ongoing efforts towards inclusivity and technological empowerment for the South African DHH community.
Included in
Computer and Systems Architecture Commons, Digital Communications and Networking Commons
Ubuntu Signs: A Tool for Real-time Translation of South African Sign Language
Kumasi, Ghana & Virtual
Deaf and Hard-of-Hearing (DHH) individuals often face communication challenges with people who do not understand sign language. Existing systems lack support for South African Sign Language (SASL) and are limited in accessibility, efficiency and accuracy. This paper introduces Ubuntu Signs, a mobile prototype tool for real-time SASL translation using a camera-based interface and machine learning. The tool uses two models to recognize gestures: a static model for alphabetical and numerical gestures and a dynamic model for continuous gestures representing words and phrases. A preliminary user study evaluated the effectiveness, efficiency, accuracy and satisfaction of Ubuntu Signs. Results indicated high usability and positive user satisfaction, particularly for the tool’s simplicity, intuitive design and accurate recognition of static gestures. However, limitations were observed in dynamic gesture recognition. Future work could focus on improving this model. This study contributes to ongoing efforts towards inclusivity and technological empowerment for the South African DHH community.
