Author(s): K. Girija, R. Bhuvanya, D. Saravanan, B. Kumaragurubaran

Abstract: In the present study, we present a scheme for recognizing partial speech with the aid of nonaudible murmur (NAM). NAM is said to be a quiet murmur that is so low that even people nearby the speaker cannot hear it. It can be detected only with the help of a special body-conductive microphone termed as NAM microphone. This can help people who are in need of private conversation in a public place without annoying others. This approach can be used for impaired people who can hear sound but can speak only few partial words (Semi-mute). Our proposed method makes use of Tamil speech recognition engine (TSRE) and NAM microphone. Experimental results have proved that NAM microphone is far better in recognizing partial speech compared to normal microphone.