Feature-dependent compensation of coders in speech recognition

Author	dc.contributor.author	Becerra Yoma, Néstor
Author	dc.contributor.author	Molina, Carlos	es_CL
Admission date	dc.date.accessioned	2008-12-10T15:34:38Z
Available date	dc.date.available	2008-12-10T15:34:38Z
Publication date	dc.date.issued	2006-01
Cita de ítem	dc.identifier.citation	SIGNAL PROCESSING Volume: 86 Issue: 1 Pages: 38-49 Published: JAN 2006	en
Identifier	dc.identifier.issn	0165-1684
Identifier	dc.identifier.uri	https://repositorio.uchile.cl/handle/2250/124761
Abstract	dc.description.abstract	A solution to the problem of speech recognition with signals corrupted by coders is presented. The coding-decoding distortion is modelled as feature dependent. This model is employed to propose an unsupervised expectation-maximization (EM) estimation algorithm of the coding-decoding distortion that is able to cancel the effect of coders with as few as one adapting utterance. No knowledge about the coder is required. The feature-dependent adaptation can give a word error rate (WER) 21% lower than the feature-independent model. Finally, when compared to the baseline system, the reduction in WER can be as high as 70%.	en
Lenguage	dc.language.iso	en	en
Publisher	dc.publisher	ELSEVIER	en
Keywords	dc.subject	Speech recognition	en
Título	dc.title	Feature-dependent compensation of coders in speech recognition	en
Document type	dc.type	Artículo de revista