A hidden Markov model based keyword recognition system

Author

Rose, Richard C. ; Paul, Douglas B.

Author_Institution

Lincoln Lab., MIT, Lexington, MA, USA

fYear

1990

fDate

3-6 Apr 1990

Firstpage

129

Abstract

A speaker-independent hidden Markov model (HMM) keyword recognizer (KWR) based on a continuous-speech-recognition model is presented. The baseline keyword recognition system is described, and techniques for dealing with nonkeyword speech and linear channel effects are discussed. The training of acoustic models to provide an explicit representation of nonvocabulary speech is investigated. A likelihood ratio scoring procedure is used to account for sources of variability affecting keyword likelihood scores. An acoustic class-dependent spectral normalization procedure is used to provide explicit compensation for linear channel effects. Keyword recognition results for a standard conversational speech task with a 20-keyword vocabulary reach 82% probability of detection at a false alarm rate of 12 false alarms per keyword per hour

Keywords

Markov processes; learning systems; probability; speech analysis and processing; speech recognition; acoustic class-dependent spectral normalization; acoustic model training; compensation; continuous-speech-recognition model; false alarm rate; hidden Markov model; keyword recognition system; likelihood ratio scoring procedure; linear channel effects; nonkeyword speech; nonvocabulary speech; probability; speaker-independent; Acoustic signal detection; Databases; Dynamic programming; Hidden Markov models; Laboratories; Loudspeakers; Maximum likelihood estimation; Robustness; Speech recognition; Vocabulary;

fLanguage

English

Publisher

ieee

Conference_Titel

Acoustics, Speech, and Signal Processing, 1990. ICASSP-90., 1990 International Conference on

Conference_Location

Albuquerque, NM

ISSN

1520-6149

Type

conf

DOI

10.1109/ICASSP.1990.115555

Filename

115555