MFCC computation from magnitude spectrum of higher lag autocorrelation coefficients for robust speech recognition

There are no files associated with this record.

Title MFCC computation from magnitude spectrum of higher lag autocorrelation coefficients for robust speech recognition
Author Shannon, Ben James; Paliwal, Kuldip Kumar
Publication Title Interspeech 2004 (ICSLP)
Editor Soon Hyob Kim and Dae Hee Youn
Year Published 2004
Place of publication Korea
Publisher Sunjin Printing Co.
Abstract Processing of the speech signal in the autocorrelation domain in the context of robust feature extraction is based on the following two properties: 1) pole preserving property (the poles of a given (original) signal are preserved in its autocorrelation function), and 2) noise separation property (the autocorrelation function of a noise signal is confined to lower lags, while the speech signal contribution is spread over all the lags in the autocorrelation function, thus providing a way to eliminate noise by discarding lower-lag autocorrelation coefficients). In this paper, we use these properties to derive robust features for automatic speech recognition. We compute the magnitude spectrum of the one-sided higher-lag autocorrelation sequence, process it through a Mel filter bank and parameterise it in terms of Mel Frequency Cepstral Coefficients (MFCCs). Since the proposed method combines autocorrelation domain processing with Mel filter bank analysis, we call the resulting MFCCs, Autocorrelation Mel Frequency Cepstral Coefficients (AMFCCs). Recognition experiments are conducted on the Aurora II database and it is found that the AMFCC representation performs as well as the MFCC representation in clean conditions and provides more robust performance in the presence of background noise.
Peer Reviewed Yes
Published Yes
Publisher URI
Alternative URI
ISBN 1225-441X
Conference name 8th International Conference on Spoken Language Processing
Location Jeju Island, Korea
Date From 2004-10-04
Date To 2004-10-08
Date Accessioned 2005-03-31
Language en_AU
Research Centre Institute for Integrated and Intelligent Systems
Faculty Faculty of Engineering and Information Technology
Subject PRE2009-Speech Recognition
Publication Type Conference Publications (Full Written Paper - Refereed)
Publication Type Code e1

Show simple item record

Griffith University copyright notice