Catalogo Articoli (Spogli Riviste)

OPAC HELP

Titolo:
A new approach to designing a feature extractor in speaker identification based on discriminative feature extraction
Autore:
Miyajima, C; Watanabe, H; Tokuda, K; Kitamura, T; Katagiri, S;
Indirizzi:
Nagoya Inst Technol, Dept Comp Sci, Showa Ku, Nagoya, Aichi 4668555, JapanNagoya Inst Technol Nagoya Aichi Japan 4668555 goya, Aichi 4668555, Japan ATR, Human Informat Proc Res Labs, Kyoto 6190288, Japan ATR Kyoto Japan 6190288 man Informat Proc Res Labs, Kyoto 6190288, Japan NTT, Commun Sci Labs, Kyoto 6190237, Japan NTT Kyoto Japan 6190237NTT, Commun Sci Labs, Kyoto 6190237, Japan
Titolo Testata:
SPEECH COMMUNICATION
fascicolo: 3-4, volume: 35, anno: 2001,
pagine: 203 - 218
SICI:
0167-6393(200110)35:3-4<203:ANATDA>2.0.ZU;2-W
Fonte:
ISI
Lingua:
ENG
Soggetto:
RECOGNITION; CLASSIFICATION; ALGORITHMS;
Keywords:
speaker identification; discriminative feature extraction; minimum classification error; mel-cepstral estimation; second-order all-pass system function;
Tipo documento:
Article
Natura:
Periodico
Settore Disciplinare:
Engineering, Computing & Technology
Citazioni:
48
Recensione:
Indirizzi per estratti:
Indirizzo: Miyajima, C Nagoya Inst Technol, Dept Comp Sci, Showa Ku, Gokiso Cho, Nagoya, Aichi 4668555, Japan Nagoya Inst Technol Gokiso Cho Nagoya Aichi Japan 4668555 apan
Citazione:
C. Miyajima et al., "A new approach to designing a feature extractor in speaker identification based on discriminative feature extraction", SPEECH COMM, 35(3-4), 2001, pp. 203-218

Abstract

This paper presents a new framework for designing a feature extractor in aspeaker identification system based on the discriminative feature extraction (DFE) method. In order to find the frequency scale appropriate for accurate speaker identification, a mel-cepstral estimation technique using a second-order all-pass warping function is applied to the feature extractor; the frequency warping parameters and the text-independent speaker model parameters are jointly optimized based on a minimum classification error (MCE) criterion. Experimental results show that the frequency scale after optimization is different from traditional Linear/Mel scales and the proposed system outperforms conventional systems in which only the classifier is optimized with the MCE criterion. (C) 2001 Elsevier Science B.V. All rights reserved.

ASDD Area Sistemi Dipartimentali e Documentali, Università di Bologna, Catalogo delle riviste ed altri periodici
Documento generato il 03/04/20 alle ore 14:04:06