Time-frequency distributions for automatic speech recognition

Author

Potamianos, Alexandros ; Maragos, Petros

Author_Institution

Sch. of Electr. & Comput. Eng., Georgia Inst. of Technol., Atlanta, GA, USA

Volume

9

Issue

3

fYear

2001

fDate

3/1/2001 12:00:00 AM

Firstpage

196

Lastpage

200

Abstract

The use of general time-frequency distributions as features for automatic speech recognition (ASR) is discussed in the context of hidden Markov classifiers. Short-time averages of quadratic operators, e.g., energy spectrum, generalized first spectral moments, and short-time averages of the instantaneous frequency, are compared to the standard front end features, and applied to ASR. Theoretical and experimental results indicate a close relationship among these feature sets

Keywords

Markov processes; feature extraction; spectral analysis; speech recognition; time-frequency analysis; automatic speech recognition; energy spectrum; feature sets; front end features; generalized first spectral moments; hidden Markov classifiers; instantaneous frequency; quadratic operators; short-time averages; spectral moments; time-frequency distributions; Automatic speech recognition; Band pass filters; Context modeling; Frequency estimation; Hidden Markov models; Oscillators; Speech analysis; Speech processing; Speech recognition; Time frequency analysis;

fLanguage

English

Journal_Title

Speech and Audio Processing, IEEE Transactions on

Publisher

ieee

ISSN

1063-6676

Type

jour

DOI

10.1109/89.905994

Filename

905994