Automatic Speaker Recognition Using Neural Networks and Vector Quantization

Autor: W.-C. Shieh
Kurzfassung: Finally, a combined vector quantization, multilayer perceptron and neural tree network method is proposed for speaker recognition task. A comparison of the proposed method with full search vector quantization, multilayer perceptron and neural tree network classifiers has shown that the combination method improves the speaker recognitzion results.
Dokumenttyp: Dissertation
Veröffentlichung: Universität Bremen, Bremen, Germany, Januar 1995
Index: 68
Dateien: BibTEX
Zuletzt aktualisiert am 29.04.2008 von Admin
AIT ieee GOC tzi ith Fachbereich 1
© Arbeitsbereich Nachrichtentechnik - Universität BremenImpressum / Kontakt