Frame-synchronous adaptation of cepstrum by linear regression

Author

Delphin-Poulat, Lionel ; Mokbel, Chafic

Author_Institution

CNET, Lannion, France

fYear

1997

fDate

14-17 Dec 1997

Firstpage

420

Lastpage

427

Abstract

We propose using hidden Markov models (HMMs) associated with the cepstrum coefficients to remove disturbances that degrade the speech recognition process. In order to perform this task in an online manner, we use the MUltipath Stochastic Equalization (MUSE) framework. This method allows one to process data at the frame level. Two equalization functions are examined: bias removal and linear regression. Recognition experiments carried out on both PSTN and GSM networks show the efficiency of the proposed method: thanks to MUSE, with a model trained on PSTN recorded digits, the error rate on both PSTN and GSM recorded digits can be reduced by 19% with bias subtraction and by 36% with linear regression. Similar results obtained on another vocabulary are also presented

Keywords

cepstral analysis; errors; hidden Markov models; speech recognition; statistical analysis; synchronisation; GSM networks; MUSE framework; Multipath Stochastic Equalization framework; PSTN networks; bias removal; bias subtraction; cepstrum coefficients; equalization functions; error rate; frame synchronous cepstrum adaptation; hidden Markov models; linear regression; recognition experiments; speech recognition; vocabulary; Cepstral analysis; Cepstrum; Degradation; Error analysis; GSM; Hidden Markov models; Linear regression; Speech enhancement; Speech recognition; Vocabulary;

fLanguage

English

Publisher

ieee

Conference_Titel

Automatic Speech Recognition and Understanding, 1997. Proceedings., 1997 IEEE Workshop on

Conference_Location

Santa Barbara, CA

Print_ISBN

0-7803-3698-4

Type

conf

DOI

10.1109/ASRU.1997.659119

Filename

659119