MAP and sub-word level T-norm for text-dependent speaker recognition
Entity
UAM. Departamento de Ingeniería InformáticaPublisher
International Speech Communication AssociationDate
2008-09Citation
9th Annual Conference of the International Speech Communication Association. September 22-26, 2008ISSN
2308-457XFunded by
This work was funded by the Spanish Ministry of Science and Technology under project TEC2006-13170-C02-01.Editor's Version
http://www.isca-speech.org/archive/interspeech_2008/i08_1933.htmlSubjects
speaker recognition; text-dependent; Informática; TelecomunicacionesNote
Proceedings of Interspeech 2008, Brisbane (Australia)Rights
© 2008 ISCAAbstract
This paper presents improvements in text-dependent speaker recognition based on the use of Maximum A Posteriori (MAP) adaptation of Hidden Markov Models and the use of new sub-word level T-Normalization procedures. Results on the YOHO corpus show that the use of MAP adaptation provides a relative improvement of 22.6% in Equal Error Rate (EER) in comparison with Baum-Welch retraining and Maximum Likelihood Linear Regression (MLLR) adaptation. The newly proposed sub-word level T-Normalization procedures provide additional relative improvements, particularly for small cohorts, of up to 20% in EER in comparison with the normal utterance-level T-Normalization.
Files in this item
Google Scholar:Toledano, Doroteo T.
-
Hernández López, Daniel
-
Esteve Elizalde, Cristina
-
González Rodríguez, Joaquín
-
Fernández Pozo, Rubén
-
Hernández Gómez, Luis
This item appears in the following Collection(s)
Related items
Showing items related by title, author, creator and subject.