Title :
Bangla triphone HMM based word recognition
Author :
Hasan, Mohammad Mahedi ; Hassan, Foyzul ; Islam, Gazi Md Moshfiqul ; Banik, Manoj ; Kotwal, Mohammed Rokibul Alam ; Rahman, Sharif Mohammad Musfiqur ; Muhammad, Ghulam ; Mohammad, Nurul Huda
Author_Institution :
Blueliner Bangladesh, Dhaka, Bangladesh
Abstract :
In this paper, we have prepared a medium size Bangla speech corpus and compare performances of different acoustic features for Bangla word recognition. Most of the Bangla automatic speech recognition (ASR) system uses a small number of speakers, but 40 speakers selected from a wide area of Bangladesh, where Bangla is used as a native language, are involved here. In the experiments, mel-frequency cepstral coefficients (MFCCs) are inputted to the triphone hidden Markov model (HMM) based classifiers for obtaining word recognition performance. From the experiments, it is shown that MFCC-based method of 39 dimensions provides a higher word correct rate (WCR) and word accuracy (WA) than the other methods investigated. Moreover, a higher WCR and WA is obtained by the MFCC39-based method with fewer mixture components in the HMM.
Keywords :
cepstral analysis; hidden Markov models; speech recognition; Bangla triphone HMM; MFCC39-based method; automatic speech recognition system; hidden Markov model based classifiers; medium size Bangla speech corpus; mel-frequency cepstral coefficients; word accuracy; word correct rate; word recognition; Artificial neural networks; Asia; DH-HEMTs; Hidden Markov models; Mel frequency cepstral coefficient; Speech; Speech recognition; automatic speech recognition; hidden Markov model; mel-frequency cepstral coefficients; triphone model;
Conference_Titel :
Circuits and Systems (APCCAS), 2010 IEEE Asia Pacific Conference on
Conference_Location :
Kuala Lumpur
Print_ISBN :
978-1-4244-7454-7
DOI :
10.1109/APCCAS.2010.5775010