Deprecated (16384): The ArrayAccess methods will be removed in 4.0.0.Use getParam(), getData() and getQuery() instead. - /home/brlfuser/public_html/src/Controller/ArtileDetailController.php, line: 150
 You can disable deprecation warnings by setting `Error.errorLevel` to `E_ALL & ~E_USER_DEPRECATED` in your config/app.php. [CORE/src/Core/functions.php, line 311]
Deprecated (16384): The ArrayAccess methods will be removed in 4.0.0.Use getParam(), getData() and getQuery() instead. - /home/brlfuser/public_html/src/Controller/ArtileDetailController.php, line: 151
 You can disable deprecation warnings by setting `Error.errorLevel` to `E_ALL & ~E_USER_DEPRECATED` in your config/app.php. [CORE/src/Core/functions.php, line 311]
Warning (512): Unable to emit headers. Headers sent in file=/home/brlfuser/public_html/vendor/cakephp/cakephp/src/Error/Debugger.php line=853 [CORE/src/Http/ResponseEmitter.php, line 48]
Warning (2): Cannot modify header information - headers already sent by (output started at /home/brlfuser/public_html/vendor/cakephp/cakephp/src/Error/Debugger.php:853) [CORE/src/Http/ResponseEmitter.php, line 148]
Warning (2): Cannot modify header information - headers already sent by (output started at /home/brlfuser/public_html/vendor/cakephp/cakephp/src/Error/Debugger.php:853) [CORE/src/Http/ResponseEmitter.php, line 181]
LATEST NEWS UPDATES | Indian professors developing authentication system through speech recognition by Bikash Singh & Debjoy Sengupta

Indian professors developing authentication system through speech recognition by Bikash Singh & Debjoy Sengupta

Share this article Share this article
published Published on Dec 3, 2010   modified Modified on Dec 3, 2010

Scenario I: An illiterate person walks into an ATM, utters his password in Bhojpuri to withdraw money. Or a labourer working under the National Rural Employment Guarantee Act (NREGA) speaks his voice signature to mark his attendance for the day.

Scenario II: The door at a top-secret Indian defence establishment open son a voice recognition system that allows only certain officials access to the premises.

If you think this is stuff out of movies, think again. A team of Indian professors is developing a person authentication system based on the speech verification method .

Simply put, it’s a solution that can be retrofitted into any password authentication system and will recognise your voice as a password — irrespective of the language you speak. And when launched, it is likely to work in all environments — offices, railway stations, airports or at an ATM.

By 2011 a pilot project will be rolled out with commercial use expected the year after. “We have been approached by a few companies to use the system in industrial environments as well as for its commercial production,” says Dr SR Mahadeva Prasanna, associate professor at the department of electronics and communication engineering at Indian Institute of Technology-Guwahati (IIT-G).

Prasanna leads an eight-member team which is working on the project, being funded by the department of science and technology.

He, however, says the underlying concept is not entirely new but the challenge is that available systems only recognise English, that too with UK or US accents. “The real challenges are to make the system Indian-language compatible and fail or fraud proof. Plus, it will have to support different accents of Hindi as well as Indian English. It will also have to work perfectly well in noisy environments,” the professor said.

“Indian office environments are different from that in the UK or the US. An Indian office almost is always noisy. So are places like railway stations, airports or even some ATMs, for that matter. The system being developed will cut out background noise and capture only the voice.

Morever, a person’s voice may vary and it can be different due to reasons including cold or a choked throat. We are now busy making sure that the system can handle such voice deviations and yet be fail-proof,” said Dr Rohit Sinha, who’s also working with the professor on the project and is an assistant professor at IIT-Guwahati.

The authentication system has a speech processing security application that can be used in e-commerce platforms apart from in high-security areas. The system has been tested on a group of 100 individuals speaking a mix of some 11 languages as well as different accents of Hindi and English.

It has also been tested through various voice input systems like headphones, tablet PCs, microphones, headphones and digital voice recorders in uncontrolled environments. The sample size was recently raised to 200 and the number of languages to 18. Plus, the system was tested on a mobile phone platform.

But there are sceptics. Mr Nitin Khanapurkar, executive director at KPMG and a specialist in IT and ITeS, says: “Such voice-based authentication system has not been much successful in the past. A voice can be faked much easily than any other standard biometric impressions like iris and thumb imprint. Additionally, voice modulation varies with climate and the surrounding environments making it difficult for such systems to function properly.”

“Such systems can at most be used as a supplementary or complimentary authentication system. Difficulty in India is manifold with many accents of the same language as well as large number of languages,” said Mr Khanapurkar.

Dr Prasanna, however, said: “That the team is working towards making it fake proof.” Voice, he argues, is also the biometric feature of an individual and can’t be same for two persons, hence is largely fake- proof. According to Mr Prasanna, some parts of the voice of a fraudster will differ from that of the original person — enough for the system to keep the crooks at bay.


The Economic Times, 2 December, 2010, http://economictimes.indiatimes.com/tech/ites/Indian-professors-developing-authentication-system-through-speech-recognition/articleshow/7025697.cms


Related Articles

 

Write Comments

Your email address will not be published. Required fields are marked *

*

Video Archives

Archives

share on Facebook
Twitter
RSS
Feedback
Read Later

Contact Form

Please enter security code
      Close