AI-WEAR: SMART TEXT READER FOR BLIND/VISUALLY IMPAIRED STUDENTS USING RASPBERRY PI WITH AUDIO-VISUAL CALL AND GOOGLE ASSISTANCE

Abstract

The goal of this research is to create a prototype referred to as AI-WEAR: Smart Text Reader for Blind or Visually Impaired Students, which utilizes Raspberry Pi equipped with Audio-Visual Call and Google Assistance. The prototype incorporates various functionalities including text-to-speech capability for reading, Google assistance for online support, and video streaming through Jitsi Meet, enabling students to interact with their teachers. The device offers two modes of control: voice commands and user-friendly buttons with Braille letters engraved on them. OCR (Optical Character Recognition) and Text-to-Speech are integrated into the system. Synthesis techniques on the Raspberry Pi platform. By utilizing OCR, the device scans and extracts text, which is then converted into audio output through a headset. Additionally, the device employs GSM/GPRS technology to access the internet via cellular data when Wi-Fi connectivity is unavailable. The researcher employed the Long-Short Term Memory and Image processing algorithms in this project, and extensive testing and maintenance have been conducted, resulting in favourable evaluation outcomes. The prototype has successfully met the desires of the ISO/IEC 25010 standard. Although some adjustments may be necessary, this proposed device has significant potential to provide visually impaired individuals with innovative learning opportunities, especially in distance education settings. Furthermore, a cost-comparative analysis for future mass production of this assistive prototype tool for blind and visually impaired has been conducted.