Android Based Sinhala Document Reader for Visually Impaired People

Abstract

In this era knowledge is the key factor for successful survival. Knowledge is gained through many ways and reading is one of them. For a person who is visually impaired, reading can be performed through braille system. In today’s context, this traditional braille systems are replaced with computer aided braille systems or text to speech systems. Though these technologies are widely available for English, their availability is rare for Sinhala. Also such sophisticated systems are highly priced so that an average Sri Lankan cannot afford them. Since mobile phones play a significant role in people’s life and when it comes to Sri Lanka, Android is the market leader for mobile operating systems, an Android based solution was ideal for this research problem. This research solution detects the document and automatically captures it. Then passes it to the Tesseract optical character recognition engine, which identifies text if the relevant dataset files are given and passes text output to Android’s text to speech system which reads out text input for the user using device’s default TTS engine. Also by providing necessary audio guidelines and notifications, this application makes the life easier for its users. Graphical user interface of the application is simple and designed in a way which is user friendly for visual impaired community. Functional evaluations were carried out to assess viability of the main three sub systems. Document detection accuracy is recorded as 100 percent on tested document sizes, provided document’s color and background color contrast is high. Character recognition accuracy is evaluated by comparing Sinhala and English languages and OCR’s accuracy is really low for Sinhala compared to English. Text to Speech system’s accuracy depends on OCR’s output. If accurate text inputs were given, TTS can give 100 percent accurate audio output. As a nonfunctional evaluation a usability evaluation was carried out which has scored around 90 percent positive feedback overall. Thus, this application is considered a valuable application for the visually impaired people in Sri Lanka to fulfill their craving for knowledge due to its usability and capabilities.