Presentation 2020-06-12
Improving the pronounce clarity of dysarthric speech using CycleGAN
Shuhei Imai, Takashi Nose, Aoi Kanagaki, Satoshi Watanabe, Akinori Ito,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) Several voice conversion systems have been developed that converts the dysarthric speech into healthy speech.The conventional methods, however, require a large amount of dysarthric speech for realizing a high-quality voice output.Preparing such a database is burdensome for those people with dysarthria.In this paper, we investigate a method to improve intelligibility by learning the conversion from dysarthric speech to healthy speech with multiple speakers using CycleGAN-VC2, an efficient and high-quality VC algorithm in the task of unpaired voice conversion.We trained VC models with CycleGAN-VC2 using healthy speech with multiple speaker and relatively small amount of dysarthric speech, and compared the performance of converted speech by subjective and objective evaluation.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Dysarthria / Pronounce clarity / Voice conversion / CycleGAN
Paper # WIT2020-1
Date of Issue 2020-06-05 (WIT)

Conference Information
Committee WIT
Conference Date 2020/6/12(1days)
Place (in Japanese) (See Japanese page)
Place (in English) Online meeting
Topics (in Japanese) (See Japanese page)
Topics (in English) Well-being Information Technology, etc.
Chair Daisuke Wakatsuki(Tsukuba Univ. of Tech.)
Vice Chair Shinji Sakou(Nagoya Inst. of Tech.)
Secretary Shinji Sakou(Saitama Industrial Tech. Center)
Assistant Manabi Miyagi(Tsukuba Univ. of Tech.) / Minako Hosono(AIST) / Aki Sugano(Nagoya Univ.)

Paper Information
Registration To Technical Committee on Well-being Information Technology
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Improving the pronounce clarity of dysarthric speech using CycleGAN
Sub Title (in English)
Keyword(1) Dysarthria
Keyword(2) Pronounce clarity
Keyword(3) Voice conversion
Keyword(4) CycleGAN
1st Author's Name Shuhei Imai
1st Author's Affiliation Tohoku University(Tohoku Univ.)
2nd Author's Name Takashi Nose
2nd Author's Affiliation Tohoku University(Tohoku Univ.)
3rd Author's Name Aoi Kanagaki
3rd Author's Affiliation Tohoku University(Tohoku Univ.)
4th Author's Name Satoshi Watanabe
4th Author's Affiliation Human Techno System(HTS)
5th Author's Name Akinori Ito
5th Author's Affiliation Tohoku University(Tohoku Univ.)
Date 2020-06-12
Paper # WIT2020-1
Volume (vol) vol.120
Number (no) WIT-63
Page pp.pp.1-6(WIT),
#Pages 6
Date of Issue 2020-06-05 (WIT)