Transliteration, Indian Language Text to Braille in India

INTRODUCTION

The Braille encoding system is the primary means of representing textual documents in a readable format for the visually impaired. However, due to the scarcity of Braille printed reading materials, blind people in India face a daunting task while getting formal education and securing a respectable employment opportunity. (Hamara Bandhan)

The situation worsens due to the unavailability of low-cost technological support.

The National Census of India has estimated that out of around 21.9 million disabled people in India, more than 15 million people in the country are blind. This is considered to be the highest among all other disabilities.

Three out of every five children in the 0-9 year age group have been reported to be visually impaired in India.

Due to their inability in accessing information from written text documents, the blind face significant difficulties in communicating with sighted people in public places like post offices and banks where writing is the primary mode of communication.

The advent of computer systems has opened up many avenues for the visually impaired who have benefited immensely from computer-based systems like automatic text-to-Braille translation systems and audio feedback-based virtual environments. Automatic text-to-Braille translation systems are widely available for languages like English, French, Spanish, Portuguese, and Swedish. Similarly, audio feedback-based interfaces like screen readers are available for English and other languages. These technologies have enabled the visually impaired to communicate effectively with sighted people and also harness the power of the Internet.

Several works have been done on building automatic, bi-directional text to Braille transliteration system and speech- enabled interfaces for the visually impaired community.

However, most of the systems cannot be directly used by the large visually impaired population in the Indian sub-continent due to the following:

§ Most systems are based on foreign languages like English, French, Germany , Spanish, Portuguese, and Swedish.

§ Indian language scripts are quite different from that of European or American languages, necessitating separate rules to transliterate Indian language texts to Braille.

§ Foreign systems like, Duxbury and JAWS are costly, given the Indian economic reality of the visually impaired population belonging to the poorer section.

To overcome these challenges and limitations of the existing systems, we present the Sparsha tool set, a speech- enabled transliteration system from Indian language text to Braille. Reverse transliteration: The Sparsha system allows reverse transliteration of Braille to text both for Indian languages as well as English. This allows the visually impaired to communicate seamlessly with other sighted people. The Braille code to be translated may be entered into the computer using a standard six key Braille keyboard. After translating the Braille code into text, the visually readable text may then be checked for correctness using a file reading system which will be described in later section.

In order to achieve reverse translation from Braille to text, the system uses a finite state machine based approach similar to that used for translating text to Braille as described previously. The task of reverse translation also uses the code tables corresponding to the language to which the text is being translated. Thus the system can easily be extended to other languages just by adding the corresponding code tables to achieve both forward and reverse translation.

The system provides a generic framework for the transliteration of large number of popular Indian language texts to Braille.

Further, the system can also be used as an Indian language document reader where a user can select a particular text document to get the corresponding speech output. This allows the present system to be used by both, a sighted as well as a blind person. We also extended our present system to provide Braille transliteration of Dzongkha texts, keeping in view Bhutan ’s proximity with India .

Currently, our system handles only four Indian languages: Hindi, Bengali, Assamese, and English, along with Dzongkha. However, due to the system’s unified framework, any other Indian language can be easily be added to it.

Indian Language Braille Encoding System

The Bharati Braille system is the standard technique of representing Indian language texts to Braille . The system uses six dot cells to represent each character. The combination of these six dots can generate 63 (26-1) different Braille characters.

It has been observed that for all Indian languages, the corresponding Braille code is the same.

In other words, a distinct Braille cell may correspond to a different Indian language character. However, a single Indian language character may require one or more than one Braille cell for its representation.

One of the most interesting phenomena found in Indian language scripts is the usage of composite or conjugate characters. Conjugates are often constructed by the sequential concatenation of two or more characters with a special character called halant. The construction of conjugate characters follows certain rules.

The conjugate characters, as constructed by clustering of consonants and vowels, may have an entirely different visual representation. However, the corresponding transliterated Braille is represented by a sequence of Braille cells for each of the characters.  READ MORE HERE

India’s first registered Braille newspaper in Hindi ‘Drishti’ by Reliance in Mumbai on March 19, 2012.

RELATED: Sparsha: A Comprehensive Indian Language Tool set for the Blind

REGIONAL CONFERENCE ON BRAILLE UNIFORMITY – PARIS 1950

Central America’s Braille Newspaper

BRAILLE in different languages