The principle of recognition is simple pattern matching in the parameter space with nonlinear adjustment of the time axis.
The spectrum envelope and cepstrum attain the best recognition score of 100% for ten spoken digits of a single male speaker. PAC's seem to be preferable because of their ease of extraction and theoretical orthogonalities; however, these PAC's tend to suffer from computation errors when computed by fixed-point arithmetic with a short accumulator length. We find two effective means to improve the errors; one is variable use of the PAC dimensions controlled by computation accuracy, and the other is smoothing along of time axis. With these improvements the PAC's offer almost 100% recognition.
[Speech Recognition] [Research Themes of Prof. Nakano.]
| mail address: | |
<- Please enter the string into the address field |
First Written Before June 17, 1998
Transplanted to KSU Before June 18, 2003
Transplanted to So-net April 22, 2007
Last Update April 22, 2007
© Yasuaki Nakano 1998-2007