Catalogo Articoli (Spogli Riviste)

OPAC HELP

Titolo:
Relevance of time-frequency features for phonetic and speaker-channel classification
Autore:
Yang, HH; Van Vuuren, S; Sharma, S; Hermansky, H;
Indirizzi:
Oregon Grad Inst Sci & Technol, Dept Elect & Comp Engn, Beaverton, OR 97006 USA Oregon Grad Inst Sci & Technol Beaverton OR USA 97006 erton, OR 97006 USA
Titolo Testata:
SPEECH COMMUNICATION
fascicolo: 1, volume: 31, anno: 2000,
pagine: 35 - 50
SICI:
0167-6393(200005)31:1<35:ROTFFP>2.0.ZU;2-3
Fonte:
ISI
Lingua:
ENG
Keywords:
mutual information; sources of variability; spectral feature; input selection; phonetic classification; multi-layer perceptron;
Tipo documento:
Article
Natura:
Periodico
Settore Disciplinare:
Engineering, Computing & Technology
Citazioni:
14
Recensione:
Indirizzi per estratti:
Indirizzo: Yang, HH Oregon Grad Inst Sci & Technol, Dept Elect & Comp Engn, 20000 NW Walker Rd, Beaverton, OR 97006 USA Oregon Grad Inst Sci & Technol 20000 NW Walker Rd Beaverton OR USA 97006
Citazione:
H.H. Yang et al., "Relevance of time-frequency features for phonetic and speaker-channel classification", SPEECH COMM, 31(1), 2000, pp. 35-50

Abstract

The mutual information concept is used to study the distribution of speechinformation in frequency and in time. The main focus is on the informationthat is relevant for phonetic classification. A large database of hand-labeled fluent speech is used to (a) compute the mutual information (MI) between a phonetic classification variable and one spectral feature variable in the time-frequency plane, and (b) compute the joint mutual information (JMI) between the phonetic classification variable and two feature variables inthe time-frequency plane. The MI and the JMI of the feature variables are used as relevance measures to select inputs for phonetic classifiers. Multi-layer perceptron (MLP) classifiers with one or two inputs are trained to recognize phonemes to examine the effectiveness of the input selection method based on the MI and the JMI, To analyze the non-linguistic sources of variability, we use speaker-channel labels to represent different speakers anddifferent telephone channels and estimate the MI between the speaker-channel variable and one or two feature variables. (C) 2000 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 08/04/20 alle ore 11:54:51