Smart Spectacles for The Deaf with Voice to Text and Sign Language Integration
Md. Mehedi Hassan, MD. Ashik Mahmud, Abrar Shahriyar, Naquibuddin Sarkar, Sonjoy Chandra Mohonto, Md Jakir Hossain, Golam Rakib Chowdhury · 2023
This research introduces an innovative method to facilitate communication for those with hearing impairments by using a technologically advanced eyewear device capable of converting spoken language into written text and sign language. This study examines the design and implementation of the system, considering its potential and practical implications. A comprehensive examination of existing literature and a thorough assessment of practical demands were undertaken to ascertain crucial factors and prerequisites. This study also evaluates the feasibility and impacts of the system after its deployment, while also considering elements that could potentially alleviate any potential problems. This study assesses the societal and health-related implications of the system, considering the necessary technical specifications. The dataset has been generated and refined using the software "Audacity". The voice recognition machine learning model was constructed on the "Edge Impulse" platform. The device chassis layout was created using "Autodesk Fusion 360," and a deep learning model, specifically a convolutional neural network (CNN), was produced. The evaluation of its performance yielded a result of approximately 90%, prompting the suggestion of design modifications to align with expectations while considering the intricacies and limitations of the prototype. In general, the system presents a potentially effective resolution for enhancing communication among those with hearing impairments, and the research study provides valuable perspectives on its creation and application.