Real-Time Sign Language-to-Speech Translation System With AI-Powered Wearable Technology

Globally, over 430 million people live with disabling hearing loss, a number projected to rise to 2.5 billion by 2050 according to the World Health Organization. Among them, individuals with mutism or profound hearing impairments often rely on sign language as their primary mode of communication. However, a communication gap persists, as most people in the general population are not proficient in sign language. This barrier hampers inclusion, access to services, and full societal participation for individuals with hearing and speech disabilities. To address this challenge, we present SignSpeak, a real-time cyber-physical system that translates sign language gestures into audible speech, enabling interactions between sign language users and non-signers. The system features a wearable glove embedded with flex sensors on each finger and a spatial sensor for wrist tracking. Data from these sensors is collected by an Arduino microcontroller and processed using machine learning models implemented in Python and TensorFlow to recognize and classify gestures accurately. The recognized gestures are then converted into synthesized speech output. By combining low-cost hardware with scalable software, SignSpeak offers an affordable, portable, and user-friendly solution to bridge the communication divide. We evaluate the system in terms of recognition accuracy, responsiveness, and practicality, demonstrating its potential to support real-time communication in educational, professional, and social contexts. Experimental validation confirmed that SignSpeak achieved an average recognition accuracy of 94.7% across 29 gesture classes, including alphabet letters and basic expressions, with an end-to-end response time of 112 ms. These results demonstrate both reliability and real-time usability, reinforcing the potential of the proposed solution as an accessible communication aid.