Bangla Automatic Phoneme Recognition

Abstract

Speech recognition is now known a most challenging problem. People are trying to use speech recognition in so many applications to make their work easier. In Bengali language there are so many varieties in pronunciation. They talk in their regional language.

In our thesis work we have tried to work with these different languages of Bangla. We took their voice recordings and made experiments.

We have tried to convert the recordings into text. To recognize the device with the language we tried our level best. Here we used MFCC inputted to the HMM based classifier for getting Phoneme recognition performance.