Phoneme Modeling for Speech Recognition in Kannada using Multivariate Bayesian Classifier

We build an automatic phoneme recognition system based on Bayesian Multivariate Modeling which is a static scheme. Phoneme models were built by using stochastic pattern recognition and acoustic phonetic schemes to recognise phonemes. Since our native language is Kannada, a rich South Indian Language, we have used 15 Kannada phonemes to train and test these models. As Mel – Frequency Cepstral Coefficients (MFCC) are well known acoustic features of speech, we have used the same in speech feature extraction. Finally performance analysis of models in terms of Phoneme Error Rate (PER) justifies the fact that though static modeling yields good results, improvization is necessary in order to use it in developing Automatic Speech Recognition systems. KeywordsBayesian Classification, Kannada, MFCC, Pattern Recognition; PER, Phoneme Modeling