Implementation of Phonetic Level Speech Recognition in Kannada Using HTK

This paper aims to build an automatic speech recognition system (ASR) to recognize Kannada words spoken continuously, using an open source speech recognition tool, HTK. With a focus on tourism application, a speaker independent, mid-sized vocabulary voice corpus with a dictionary size of 250 has been created for this purpose. Mel frequency cepstral coefficients (MFCC) has been used to extract the feature vectors. The system uses Hidden Markov model(HMM) with triphone acoustic modeling. The speech recognition rate is checked for offline and online recognition mode for male and female speakers.