This study proposes a new training method for hidden Markov model with separate vector quantization (SVQ-HMM) in speech recognition. The proposed method uses the correlation of two different kinds of features: cepstrum and delta-cepstrum. The correlation is used to decrease the number of reestimation for two features thus the total computation time for training models decreases. The proposed method is applied to Japanese language isolated dgit recognition.
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copy
Yasuhisa HAYASHI, Satoshi KONDO, Nobuyuki TAKASU, Akio OGIHARA, Shojiro YONEDA, "An SVQ-HMM Training Method Using Simultaneous Generative Histogram" in IEICE TRANSACTIONS on Fundamentals,
vol. E75-A, no. 7, pp. 905-907, July 1992, doi: .
Abstract: This study proposes a new training method for hidden Markov model with separate vector quantization (SVQ-HMM) in speech recognition. The proposed method uses the correlation of two different kinds of features: cepstrum and delta-cepstrum. The correlation is used to decrease the number of reestimation for two features thus the total computation time for training models decreases. The proposed method is applied to Japanese language isolated dgit recognition.
URL: https://global.ieice.org/en_transactions/fundamentals/10.1587/e75-a_7_905/_p
Copy
@ARTICLE{e75-a_7_905,
author={Yasuhisa HAYASHI, Satoshi KONDO, Nobuyuki TAKASU, Akio OGIHARA, Shojiro YONEDA, },
journal={IEICE TRANSACTIONS on Fundamentals},
title={An SVQ-HMM Training Method Using Simultaneous Generative Histogram},
year={1992},
volume={E75-A},
number={7},
pages={905-907},
abstract={This study proposes a new training method for hidden Markov model with separate vector quantization (SVQ-HMM) in speech recognition. The proposed method uses the correlation of two different kinds of features: cepstrum and delta-cepstrum. The correlation is used to decrease the number of reestimation for two features thus the total computation time for training models decreases. The proposed method is applied to Japanese language isolated dgit recognition.},
keywords={},
doi={},
ISSN={},
month={July},}
Copy
TY - JOUR
TI - An SVQ-HMM Training Method Using Simultaneous Generative Histogram
T2 - IEICE TRANSACTIONS on Fundamentals
SP - 905
EP - 907
AU - Yasuhisa HAYASHI
AU - Satoshi KONDO
AU - Nobuyuki TAKASU
AU - Akio OGIHARA
AU - Shojiro YONEDA
PY - 1992
DO -
JO - IEICE TRANSACTIONS on Fundamentals
SN -
VL - E75-A
IS - 7
JA - IEICE TRANSACTIONS on Fundamentals
Y1 - July 1992
AB - This study proposes a new training method for hidden Markov model with separate vector quantization (SVQ-HMM) in speech recognition. The proposed method uses the correlation of two different kinds of features: cepstrum and delta-cepstrum. The correlation is used to decrease the number of reestimation for two features thus the total computation time for training models decreases. The proposed method is applied to Japanese language isolated dgit recognition.
ER -