Confidence measure based unsupervised target model adaptation for speaker verification
作者： Alexandre Preti , Bertrand Ravera , François Capman , Jean-François Bonastre , Driss Matrouf
DOI:
关键词: Reduction (complexity) 、 Test data 、 Channel (digital image) 、 Artificial intelligence 、 Set (abstract data type) 、 Adaptation (computer science) 、 NIST 、 Speaker recognition 、 Speech recognition 、 Computer science 、 Pattern recognition
摘要: This paper proposes a new method for updating online the client models of speaker recognition system using test data. problem is called unsupervised adaptation. The main idea proposed approach to adapt model complete set data gathered from successive test, without deciding if belongs or an impostor. adaptation process includes weighting scheme data, based on posteriori probability that targeted model. evaluated within framework NIST 2005 and 2006 Speaker Recognition Evaluations. links between channel mismatch factors also explored, both Feature Mapping Latent Factor Analysis (LFA) methods. outperforms baseline system, with relative DCF improvement 27% (37% EER). When LFA compensation technique used, achieves reduction in 20% (12.5% Index Terms: verification,
isca-speech.org
univ-avignon.fr参考文章(11)
Jean-François Bonastre, Alexandre Preti, Unsupervised model adaptation for speaker verification conference of the international speech communication association. ,(2006)
Corinne Fredouille, Jean-François Bonastre, Téva Merlin, Bayesian bpproach based decision in speaker verification. Odyssey. pp. 77- 81 ,(2001)
Eric Hansen, Raymond Slyh, Timothy Anderson, Supervised and Unsupervised Speaker Adaptation in the NIST 2005 Speaker Recognition Evaluation 2006 IEEE Odyssey - The Speaker and Language Recognition Workshop. pp. 1- 8 ,(2006) , 10.1109/ODYSSEY.2006.248122
Claudio Vair, Daniele Colibro, Fabio Castaldo, Emanuele Dalmasso, Pietro Laface, Channel Factors Compensation in Model and Feature Domain for Speaker Recognition 2006 IEEE Odyssey - The Speaker and Language Recognition Workshop. pp. 1- 6 ,(2006) , 10.1109/ODYSSEY.2006.248117
J.-F. Bonastre, F. Wils, S. Meignier, ALIZE, a free toolkit for speaker recognition international conference on acoustics, speech, and signal processing. ,vol. 1, pp. 737- 740 ,(2005) , 10.1109/ICASSP.2005.1415219
David A. van Leeuwen, Speaker adaptation in the NIST Speaker Recognition Evaluation 2004 conference of the international speech communication association. pp. 1981- 1984 ,(2005)
D.A. Reynolds, Channel robust speaker verification via feature mapping international conference on acoustics, speech, and signal processing. ,vol. 2, pp. 53- 56 ,(2003) , 10.1109/ICASSP.2003.1202292
P. Kenny, G. Boulianne, P. Ouellet, P. Dumouchel, Improvements in Factor Analysis Based Speaker Verification international conference on acoustics, speech, and signal processing. ,vol. 1, pp. 113- 116 ,(2006) , 10.1109/ICASSP.2006.1659970
W.M. Campbell, D.E Sturim, D.A. Reynolds, Support vector machines using GMM supervectors for speaker verification IEEE Signal Processing Letters. ,vol. 13, pp. 308- 311 ,(2006) , 10.1109/LSP.2006.870086
Claude Barras, Sylvain Meignier, Jean-Luc Gauvain, Unsupervised online adaptation for speaker verification over the telephone. The Speaker and Language Recognition Workshop (Odyssey 2004). pp. 157- 160 ,(2004)