Home » Projects » Detext: Real-Time Speech Recognition for the Deaf
Detext: Real-Time Speech Recognition for the Deaf
Team: Reyga KUY
Project Details
Project Description: Detext is an innovative, wearable device designed to assist deaf and hard-of-hearing individuals in real-time conversations. Built as a bracelet with a 16x2 LCD display and powered by an Arduino Nano, Detext provides an affordable and compact solution for overcoming communication barriers in public interactions.
Unlike traditional speech-to-text apps that require smartphones and manual operation, Detext offers a hands-free experience, allowing users to receive live transcriptions directly on their wrist. One of its key features is the intelligent punctuation system—it automatically adds question marks at the end of sentences containin"what, when, who, where, and how" since many speech recognition systems struggle to detect questions accurately.
Designed with accessibility and affordability in mind, Detext is a budget-friendly alternative to existing assistive technologies, making it more accessible to individuals with financial constraints. Its lightweight and ergonomic design ensures ease of use and comfort, while its low power consumption enhances battery efficiency (future versions may include battery integration).
While still in the prototype stage, Detext has the potential to evolve further with enhancements such as wireless connectivity, improved display options, and multi-language support. With its simplicity, effectiveness, and affordability, Detext aims to empower the deaf community by making everyday conversations more seamless and inclusive.
