Presentation	1994/5/19 Speaker-Independent Speech Recoguition Tatsuo Matsuoka,
PDF Download Page	PDF download Page Link
Abstract(in Japanese)	(See Japanese page)
Abstract(in English)	In order to apply speech recognition to practical applications such as telephone network services,recognizers should attain high recognition performance for number of unknown speakers(users).For speaker-independent speech recognition,the main issue is how to attain high recognition coping with variation of spectral characteristics which is due to the large number of speakers.In this paper,some methods for coping with speaker variation are described,in particular the methods using HMMs.The methods include the dynamic spectral feature parameter,the multiple-template method,the speaker cluster method,discriminative training,and speaker adaptation.
Keyword(in Japanese)	(See Japanese page)
Keyword(in English)	HMM / Maltiple-template / Parameter Estimation / Acoustic model / Speech database
Paper #	SP94-4
Date of Issue

Conference Information
Committee	SP
Conference Date	1994/5/19(1days)
Place (in Japanese)	(See Japanese page)
Place (in English)
Topics (in Japanese)	(See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To	Speech (SP)
Language	JPN
Title (in Japanese)	(See Japanese page)
Sub Title (in Japanese)	(See Japanese page)
Title (in English)	Speaker-Independent Speech Recoguition
Sub Title (in English)
Keyword(1)	HMM
Keyword(2)	Maltiple-template
Keyword(3)	Parameter Estimation
Keyword(4)	Acoustic model
Keyword(5)	Speech database
1st Author's Name	Tatsuo Matsuoka
1st Author's Affiliation	NTT Human Interface Laboratories()
Date	1994/5/19
Paper #	SP94-4
Volume (vol)	vol.94
Number (no)	42
Page	pp.pp.-
#Pages	8
Date of Issue