Improved classification of acoustic features via primal weight vectors

Author

Lawson, Aaron ; Harris, David ; Battles, Brandon

Author_Institution

RADC Inc., Rome, NY, USA

fYear

2011

fDate

16-19 Oct. 2011

Firstpage

89

Lastpage

92

Abstract

This paper presents a new variation of the Support Vector Machine (SVM) technique for speaker identification in audio. Primal Weight Vectors (PWV) provide a discriminative framework to distinguish between SVM models. Here we discriminate between Gaussian Mixture Model Super Vector (GSV) models, which represent one of the leading SVM based approaches to speaker identification. The PWV-GSV combination has demonstrated a consistent performance advantage over the state of the art GSV classifier on a variety of conditions.

Keywords

Gaussian processes; acoustic signal processing; feature extraction; signal classification; speaker recognition; support vector machines; Gaussian mixture model super vector model; PWV framework; PWV-GSV classifier; SVM technique; acoustic feature classification; audio feature classification; primal weight vectors; speaker identification; support vector machine technique; Acoustics; Adaptation models; Kernel; Speaker recognition; Support vector machine classification; Vectors; acoustic features; primal weight vectors; speaker recognition; support vector machines;

fLanguage

English

Publisher

ieee

Conference_Titel

Applications of Signal Processing to Audio and Acoustics (WASPAA), 2011 IEEE Workshop on

Conference_Location

New Paltz, NY

ISSN

1931-1168

Print_ISBN

978-1-4577-0692-9

Electronic_ISBN

1931-1168

Type

conf

DOI

10.1109/ASPAA.2011.6082346

Filename

6082346