DocumentCode :
2704665
Title :
Discriminative Vector for Spoken Language Recognition
Author :
Bin Ma ; Rong Tong ; Haizhou Li
Author_Institution :
Inst. for Infocomm Res., Singapore
Volume :
4
fYear :
2007
fDate :
15-20 April 2007
Abstract :
We propose a language recognition system based on discriminative vectors, in which parallel phone recognizers serve as the voice tokenization front-end followed by vector space modeling that effectively vectorizes phonotactic features, and the final classification is carried out based on the discriminative vectors. We design an ensemble of discriminative binary classifiers. The output values of these classifiers construct a discriminative vector, also referred to as output codes, to represent the high-dimensional phonotactic features. We achieve equal-error-rate of 1.95%, 3.02% and 4.9% on 1996, 2003 and 2005 NIST LRE databases, respectively, for 30-second trials.
Keywords :
speech recognition; discriminative binary classifiers; discriminative vector; parallel phone recognizers; phonotactic features; spoken language recognition; Artificial neural networks; Feature extraction; NIST; Natural languages; Principal component analysis; Spatial databases; Speech recognition; Statistics; Support vector machine classification; Support vector machines; discriminative vector; ensemble classifiers; output codes; spoken language recognition;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Acoustics, Speech and Signal Processing, 2007. ICASSP 2007. IEEE International Conference on
Conference_Location :
Honolulu, HI
ISSN :
1520-6149
Print_ISBN :
1-4244-0727-3
Type :
conf
DOI :
10.1109/ICASSP.2007.367241
Filename :
4218272
Link To Document :
بازگشت