DocumentCode
806863
Title
Feature compensation based on switching linear dynamic model
Author
Kim, Nam Soo ; Lim, Woohyung ; Stern, Richard M.
Author_Institution
Sch. of Electr. Eng., Seoul Nat. Univ., South Korea
Volume
12
Issue
6
fYear
2005
fDate
6/1/2005 12:00:00 AM
Firstpage
473
Lastpage
476
Abstract
In this letter, we propose a novel approach to feature compensation for robust speech recognition in noisy environments. We employ the switching linear dynamic model (SLDM) as a parametric model for the clean speech distribution, which enables us to exploit temporal correlations inherent in speech signals. Both the background noise and clean speech components are simultaneously estimated by means of the interacting multiple model (IMM) algorithm.
Keywords
correlation theory; speech recognition; IMM algorithm; SLDM; clean speech component; feature compensation; interacting multiple model; noisy environment; speech recognition; switching linear dynamic model; Acoustic noise; Background noise; Degradation; Helium; Noise robustness; Parametric statistics; Speech enhancement; Speech recognition; Vectors; Working environment noise; Feature compensation; robust speech recognition; switching linear dynamic model (SLDM);
fLanguage
English
Journal_Title
Signal Processing Letters, IEEE
Publisher
ieee
ISSN
1070-9908
Type
jour
DOI
10.1109/LSP.2005.847862
Filename
1430750
Link To Document