Presentation 2014-12-16
Accent identification by conbining GMM and DNN under reverberant environment
Ryota Sakagami, Longbiao Wang, Zhang Zhaofeng, Khomdet Phapatanaburi, Masahiro Iwahashi,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) Speech recognition technical is becoming popular in many fields, while the acoustic model for native is not effective for non-native speakers. Thus, we need to distinguish the accent of native and non-native speech and select the appropriate acoustic model for speech recognition. In this paper, we consider the effect of reverberation in reality application. Thus, two methods of accent verification by using GMM and DNN in reverberant environment are proposed. The combination of likelihood with these two approaches is also performed. In reverberant environment, we improved the accent verification rate from 49.9% to 90.7% with GMM and 90.4% with DNN. Furthermore, the combination of likelihood gives a better verification rate of 96.6%.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) machine learning / GMM / DNN / accent verification
Paper # SP2014-120
Date of Issue

Conference Information
Committee SP
Conference Date 2014/12/8(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) Accent identification by conbining GMM and DNN under reverberant environment
Sub Title (in English)
Keyword(1) machine learning
Keyword(2) GMM
Keyword(3) DNN
Keyword(4) accent verification
1st Author's Name Ryota Sakagami
1st Author's Affiliation Nagaoka University of Technology()
2nd Author's Name Longbiao Wang
2nd Author's Affiliation Nagaoka University of Technology
3rd Author's Name Zhang Zhaofeng
3rd Author's Affiliation Nagaoka University of Technology
4th Author's Name Khomdet Phapatanaburi
4th Author's Affiliation Nagaoka University of Technology
5th Author's Name Masahiro Iwahashi
5th Author's Affiliation Nagaoka University of Technology
Date 2014-12-16
Paper # SP2014-120
Volume (vol) vol.114
Number (no) 365
Page pp.pp.-
#Pages 6
Date of Issue