Centre for Development of Advanced Computing (C-DAC) and IBM India Research Laboratory today announced the development of Shrutlekhan-Rajbhasha, a Hindi speaker independent, continuous speech recognition system using IBM Desktop Hindi Speech Recognition technology.

Shrutlekhan-Rajbhasha is part of C-DAC’s ongoing MANTRA-Rajbhasha project for ensuring a high level of accuracy in English to Hindi translation in the domain of Administration, Finance, Agriculture and Small Scale Industry sponsored by Department of Official Language, Ministry of Home Affairs, Government of India. The system is used as a tool in the MANTRA-Rajbhasha System to facilitate the correction of the translated Hindi text.

The IBM Desktop Hindi Speech Recognition technology developed by IBM India Research Laboratory has been used to integrate the speaker independent continuous speech recognition of Hindi to edit the translated text through Shrutlekhan-Rajbhasha in the MANTRA-Rajbhasha System.

The IBM Desktop Hindi Speech Recognition technology understands and transcribes human speech with little use of keyboards, thereby helping people unfamiliar with computers or the Hindi language. Since there are various keyboards in use for Indian languages, speech recognition eliminates the need to learn different keyboard mapping. The technology has been tested and trained for variations over a large number of speakers from different regions of the country. The technology could be also useful in voice-enabling ATM kiosks and in car navigation systems.

In the Indian context, there are large variations in the pronunciation of the same word by various users hailing from different parts of the country. C-DAC has also collected a large amount of data from users belonging to different states and captured the variations in speech output. A data analysis was conducted to build the spellchecker for Hindi to correct the spoken-word-errors that has increased the accuracy of the system and it also includes the domain specific dictionaries to provide a better quality output.

Shrutlekhan-Rajbhasha has also been integrated with a number of other user-friendly features such as the facility for conversion of Text to Number (digits & decimal), Date & Currency format. It also provides the facility to convert Unicode text into ISFOC fonts to enable a user to import the text to any Windows based application.

“Shrutlekhan-Rajbhasha, developed by C-DAC for promoting the use of Hindi for official and other purposes, is one example of how IBM speech recognition technology might be used to make people’s lives easier” said Dr. Daniel Dias, Director, IBM India Research Laboratory. “IBM believes collaborations such as this hold the key to driving innovation that matters for the world.”

Shri S Ramakrishnan, Director General, C-DAC added, “the coming together of IBM and C-DAC is a step forward towards enabling speech technologies to benefit sizeable Indian population already using Hindi to benefit from a greater degree of technology penetration in their daily lives. Technological usefulness is largely determined by the modes of accessibility it offers, and speech technologies can play a vital role towards paving the way for its acceptance, even beyond the limitations of literacy.”

“C-DAC as the pioneers of Indian language computing in the country, is happy to combine its expertise and strengths with IBM, a global IT major for a cause that is both far reaching in expanse and noble”, said Shri Ramakrishnan.

“As technologies, markets and social conventions have evolved over years, IBM Research has adjusted its methods for stimulating innovation. This has strengthened its high-end collaborative work processes. This is significant since isolation inhibits innovation. We are proud of our special talent pool that works in close association with the end-users - the strategic partners of innovation. We are committed to sustain a culture of innovation, and, we will apply our discoveries and insights to important business and societal issues,” said Dr. Dias.