VoiceBridge: An Adaptive Bilingual Speech Therapy Support Application Using Context-Aware Deep Neural Networks

Juliet Arthur, Prince Dawson Tetteh, Dickson Marfo Fosu

Keywords: Assistive Artificial Intelligence, Speech Recognition, Deep Neural Networks, Adaptive Learning Systems, Bilingual Speech Technology

TL;DR: VoiceBridge is a bilingual (English–Twi) mobile application that uses deep neural networks and adaptive learning to support speech therapy and improve pronunciation for individuals with speech difficulties in low-resource settings.

Abstract: Speech impairments significantly affect communication, educational participation, and social inclusion, particularly in low-resource settings where access to professional speech therapists is limited. This work presents VoiceBridge, a bilingual mobile application designed to assist individuals with speech difficulties through adaptive, deep neural network techniques. The system supports English and Twi languages and follows structured speech therapy methodologies, beginning with phonemic awareness using alphabets and numbers and progressively advancing to syllables and contextual speech exercises. The application processes user speech input using acoustic feature extraction techniques and evaluates pronunciation accuracy through a deep neural network‑based speech classification model trained on reference phoneme patterns.

Email Sharing: We authorize the sharing of all author emails with Program Chairs.

Data Release: We authorize the release of our submission and author names to the public in the event of acceptance.

Submission Number: 3