Bidirectional Braille Transcription for Kannada and Telugu text using Natural Language Processing

Abstract

In today's modern society, where information is readily accessible through various sources such as the internet and newspapers, individuals with visual impairments encounter significant challenges in accessing this wealth of knowledge. Unlike their sighted counterparts who effortlessly stay informed about current events and knowledge, visually impaired individuals face obstacles in harnessing this information. To address this disparity, there is an urgent need to develop a system that enables the bidirectional conversion of natural language text into Braille, thereby offering enhanced learning opportunities for the visually impaired. This paper presents a pioneering approach to bidirectional Braille transcription for Kannada and Telugu texts, employing advanced Natural Language Processing (NLP) techniques exclusively on text-based data. Given the essential role of Braille transcription in enabling visually impaired individuals to access text, the complexity of Indian scripts like Kannada and Telugu poses unique challenges. Our proposed system utilizes state-of-the-art NLP algorithms to facilitate accurate and efficient translation between printed text and Braille. The methodology encompasses tailored preprocessing steps addressing the intricate orthographic structures of Kannada and Telugu, alongside a robust transliteration engine for converting text to Braille, and an inverse transcription mechanism to revert Braille back to standard text. Through comprehensive testing on diverse text samples, the system demonstrates high accuracy and reliability. This research significantly enhances accessibility for visually impaired Kannada and Telugu speakers and sets a precedent for the application of advanced NLP techniques in regional language Braille transcription.