DocumentCode
959925
Title
Feature Compensation Incorporating Modeling Error Statistics
Author
Lim, Woohyung ; Kim, Nam Soo
Author_Institution
Seoul Nat. Univ., Seoul
Volume
14
Issue
7
fYear
2007
fDate
7/1/2007 12:00:00 AM
Firstpage
492
Lastpage
495
Abstract
In this letter, we propose a novel approach to feature compensation for robust speech recognition in noisy environments. We analyze the error distribution of speech corruption model in the log spectral domain and represent the statistics as functions with respect to the signal-to-noise ratio. The proposed algorithm incorporates modeling error statistics into the interacting multiple model technique and shows a performance improvement over the AURORA2 speech recognition task.
Keywords
error statistics; feature extraction; speech recognition; AURORA2; error distribution; feature compensation; log spectral domain; modeling error statistics; noisy environments; robust speech recognition; signal-to-noise ratio; speech corruption model; Background noise; Discrete Fourier transforms; Error analysis; Robustness; Signal analysis; Signal to noise ratio; Speech analysis; Speech enhancement; Speech recognition; Working environment noise; Feature compensation; modeling error statistics; robust speech recognition;
fLanguage
English
Journal_Title
Signal Processing Letters, IEEE
Publisher
ieee
ISSN
1070-9908
Type
jour
DOI
10.1109/LSP.2006.891334
Filename
4244483
Link To Document