Paper Info
Title
Band-Independent Mask Estimation For Missing-Feature Reconstruction In The Presence Of Unknown Background Noise
Abstract
An effective mask estimation scheme for missing-feature reconstruction is described that achieves robust speech recognition in the presence of unknown noise. In previous work on Bayesian classification for mask estimation, white noise and colored noise were used for training mask estimators. This paper, which is concerned with both the simulation of a more diverse set of background environments and with mitigating the "sparse training" problem, describes a new Bayesian mask-estimation procedure in which each frequency band is trained independently. The new method employs colored noise for training, which is obtained by partitioning each frequency subband. We also propose a reevaluation method of voiced/unvoiced decisions to alleviate performance degradation that is caused by errors in pitch detection. Experimental results indicate that the proposed procedure in conjunction with cluster-based missing-feature imputation improves speech recognition accuracy on the Aurora 2.0 database in the presence for all types of background noise considered.
Year
DOI
Venue
2006
10.1109/ICASSP.2006.1660018
2006 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING, VOLS 1-13
Keywords
Field
DocType
background noise,frequency diversity,pitch detection,databases,feature extraction,bayesian classification,bayesian methods,white noise,speech recognition,colored noise,degradation
Value noise,Colors of noise,Background noise,Pattern recognition,Noise measurement,Computer science,White noise,Feature extraction,Speech recognition,Artificial intelligence,Pitch detection algorithm,Estimator
Conference
ISSN
Citations 
PageRank 
1520-6149
13
0.71
References 
Authors
7
2
Name
Order
Citations
PageRank
Wooil Kim112016.95
Richard M. Stern21663406.79