Presentation 1999/12/20
VocaI Tract Length Normalization using Rapid Maximum-Likelihood Estimation for Speech Recognition
Tadashi EMORI, Koichi SHINODA,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) In recent works, vocal tract length normalization methods which achieve a remapping of the frequency axis using warping functions have been proposed for a large vocabulary speech recognition system. In this work, we introduce an estimation method of the parameter characterizing individual speakers, using the remapping of the frequency axis in cepstrum domain derived from all-pass transforms. In Japanese 5000-word task speech recognition experiments, we report reductions in word error rate of 7.1% absolute. When the normalization method is combined with CMN, word error rate reduction is 14.6%.
Keyword(in Japanese) (See Japanese page)
Keyword(in English)
Paper # NLC99-101
Date of Issue

Conference Information
Committee NLC
Conference Date 1999/12/20(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 Natural Language Understanding and Models of Communication (NLC)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) VocaI Tract Length Normalization using Rapid Maximum-Likelihood Estimation for Speech Recognition
Sub Title (in English)
Keyword(1)
1st Author's Name Tadashi EMORI
1st Author's Affiliation ()
2nd Author's Name Koichi SHINODA
2nd Author's Affiliation
Date 1999/12/20
Paper # NLC99-101
Volume (vol) vol.99
Number (no) 523
Page pp.pp.-
#Pages 6
Date of Issue