Page 655 - AI for Good Innovate for Impact

AI for Good Innovate for Impact

Use case- 8: Little Signs 4.7: Education

Country: Zimbabwe

Organization: Sign<Tech>

Contact Person(s):
Jamain Vimbanashe Chindoore (tatendalerrical@ gmail .com)
Cresentia B Moyo (moyobcresentia@ gmail .com)

1 Use Case Summary Table

Category Education
The problem to be The development of a real-time speech-to-sign-language translator
addressed addresses critical communication barriers faced by the deaf and hard-of-
hearing communities. These barriers hinder access to vital information and
limit participation in social, educational, and professional environments. A
further challenge lies in ensuring that such a system can accurately translate
not only the linguistic content but also the emotional undertones, including
manipulative tactics like emotional blackmail, that might arise in conver-
sations. Emotional blackmail, which leverages fear, obligation, and guilt
to coerce behavior, must be identified and appropriately conveyed in the
translation process to empower users and ensure emotional integrity in
communication. This highlights the need for advanced natural language
processing and gesture-recognition technologies capable of capturing
both literal meaning and emotional nuance.

Key aspects of the Inclusive Education Technology
solution • Speech recognition modules (speech is converted to text via the use of
a an AI deep learning model through Google Speech API)
• Natural Language Processing (this aspect include the processing and
understanding of grammar, intent, and sentence structure which is in
alignment with the Zimbabwe sign language. In other words, deaf culture
using GPT for text parsing and translation into sign compatible struc-
tures.
• Sign Language Mapping Engine (this matches translated text to correct
sign gestures )
• Gesture and Avatar Animation Module (this displays the signs visually
via a 3D avatar )
• Zimbabwean Sign Language Gesture Database
• Real-time processing system
• • Accessibility & Localization Layer (using AI to customize translation to
local sign language and cultural context)

619