READ2U: A text-to-speech mobile application for visually impaired people

Abstract

Visually impaired people typically use Braille reading and writing system to access information. However, most textual knowledge and information are not built for Braille. To help those with visual disabilities be able to access non-Braille information, particularly in a physical format, this project presents a text-to-speech mobile application, named READ2U, which can be used to read hardcopy text to users. With this application, after a user takes a picture of a text document, the application utilizes Tesseract OCR library to analyze and extract text from the image into individual words. Then, it sends the extracted words to Text-to-Speech (TTS) system for reading them up to the user. According to the experimental results on 30 textual paragraphs with 100 words each, the accuracy of extracted Thai-only text is 73.87%, English-only text 81.27%, and Thai-English text 78.27%, respectively.