Abstract
This paper presents CRIM hidden Markov model based keyword recognition system. While there is much to say about the problem of keyword spotting and much to discuss about the keyword spotter implementation issues, the paper’s main focus is on investigating feature normalization techniques for dealing with the acoustical variabilities caused by changes in background noise, recording materials, and speaking styles. Among the techniques that are proposed and discussed are spectral subtraction for compensating noise effects and RASTA filtering and channel equalization for undoing linear channel effects. Also, techniques based on hierarchical clustering are considered for correcting data. For the Road Rally word spotting task, CRIM system’s best performance corresponds to a 75% spotting rate at 3.5 false alarms per keyword per hour.
This is a preview of subscription content, log in via an institution.
Buying options
Tax calculation will be finalised at checkout
Purchases are for personal use only
Learn about institutional subscriptionsPreview
Unable to display preview. Download preview PDF.
References
H. Hermansky and N. Morgan, “Towards Handling the Acoustic Environment in Spoken Language Processing”, Proc, 1CSLP-92, pp. 85–88, Alberta, Canada, October 1992
H. G. Hirsch, P. Meyer and H. W. Ruehl, “Improved speech recognition using high-pass filtering of subband envelopes”, Proc. EUROSPEECH-91, pp. 413–416, Genova, September 1991
D. Van Compernolle, “Noise adaptation in a hidden Markov model speech recognition system”, Computer Speech and Language 3, pp. 151–167, 1989
S. F. Boil, “Suppression of acoustic noise in speech using spectral subtraction”, IEEE Trans. Acoust., Speech, and Signal Processing, Vol. ASSP-27, No. 2, pp. 113–120, April 1979
S. Furui, “Unsupervised speaker adaptation method based on hierarchical spectral clustering”, Proc. 1CASSP-89, pp. 286–289, 1989
H. M. Cung and Y. Normandin, “Noise adaptation algorithms for robust speech recognition”, Proc. of the ESCA Tutorial and Research Workshop on Speech Processing in Adverse Conditions, pp. 171–174, Cannes-Mandelieu, November 1992
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 1995 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Cung, H.M. (1995). CRIM Hidden Markov Model Based Keyword Recognition System. In: Ayuso, A.J.R., Soler, J.M.L. (eds) Speech Recognition and Coding. NATO ASI Series, vol 147. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-57745-1_37
Download citation
DOI: https://doi.org/10.1007/978-3-642-57745-1_37
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-642-63344-7
Online ISBN: 978-3-642-57745-1
eBook Packages: Springer Book Archive