Using a Vocal Tract Length Related Parameter for Speaker Recognition

Author

Grashey, Stephan ; Geissler, C.

Author_Institution

Corporate Technol., Siemens AG, Munich

fYear

2006

fDate

28-30 June 2006

Firstpage

1

Lastpage

5

Abstract

The use of a parameter related to the vocal tract length of humans is investigated for a speaker recognition system. This parameter is a warping factor for the frequency spectrum and it was determined from a vocal tract length normalization module of an automatic speech recognizer. A decision fusion approach WHS used to combine the output of the stand-alone speaker verification with a separate classifier for the warping factor. The fused system exhibits a relative improvement of 23% compared to the stand-alone system

Keywords

decision theory; feature extraction; pattern classification; speaker recognition; speech processing; WHS; automatic speech recognizer; decision fusion approach; frequency spectrum; speaker recognition system; stand-alone speaker verification; vocal tract length; warping factor; Automatic speech recognition; Cepstral analysis; Discrete cosine transforms; Feature extraction; Finite impulse response filter; Forensics; Humans; Linear predictive coding; Mel frequency cepstral coefficient; Speaker recognition;

fLanguage

English

Publisher

ieee

Conference_Titel

Speaker and Language Recognition Workshop, 2006. IEEE Odyssey 2006: The

Conference_Location

San Juan

Print_ISBN

1-424400471-1

Electronic_ISBN

1-4244-0472-X

Type

conf

DOI

10.1109/ODYSSEY.2006.248086

Filename

4013503