IEICE Technical Committee Submission System
Conference Paper's Information
Online Proceedings
[Sign in]
Tech. Rep. Archives
 Go Top Page Go Previous   [Japanese] / [English] 

Paper Abstract and Keywords
Presentation 2023-03-01 11:00
Analysis of Noisy-target Training for DNN-based speech enhancement and investigation towards its practical use
Takuya Fujimura, Tomoki Toda (Nagoya Univ.) EA2022-112 SIP2022-156 SP2022-76
Abstract (in Japanese) (See Japanese page) 
(in English) Deep neural network (DNN)-based speech enhancement usually uses a clean speech as a training target. However, it is hard to collect large amounts of clean speech because its recording is very costly. To relax this limitation, we proposed Noisy-target Training (NyTT) that utilizes noisy speech as a training target. It has been experimentally shown that NyTT can train a DNN without clean speech. However, sufficient investigations have not been conducted to clarify the reason why NyTT works, its detailed property, and the effectiveness of utilizing large amounts of noisy speech. In this paper, we conduct various analyses to deepen our understanding of NyTT. Based on the property of NyTT, we also propose a refined method that performs higher-quality speech enhancement. Furthermore, we investigate whether using a huge amount of noisy speech is effective for improving speech enhancement performance.
Keyword (in Japanese) (See Japanese page) 
(in English) Single channel speech enhancement / Deep Neural Network / Unsupervised learning / Behavior analysis / / / /  
Reference Info. IEICE Tech. Rep., vol. 122, no. 387, EA2022-112, pp. 221-226, Feb. 2023.
Paper # EA2022-112 
Date of Issue 2023-02-21 (EA, SIP, SP) 
ISSN Online edition: ISSN 2432-6380
Copyright
and
reproduction
All rights are reserved and no part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher. Notwithstanding, instructors are permitted to photocopy isolated articles for noncommercial classroom use without fee. (License No.: 10GA0019/12GB0052/13GB0056/17GB0034/18GB0034)
Download PDF EA2022-112 SIP2022-156 SP2022-76

Conference Information
Committee SP IPSJ-SLP EA SIP  
Conference Date 2023-02-28 - 2023-03-01 
Place (in Japanese) (See Japanese page) 
Place (in English)  
Topics (in Japanese) (See Japanese page) 
Topics (in English)  
Paper Information
Registration To EA 
Conference Code 2023-02-SP-SLP-EA-SIP 
Language Japanese 
Title (in Japanese) (See Japanese page) 
Sub Title (in Japanese) (See Japanese page) 
Title (in English) Analysis of Noisy-target Training for DNN-based speech enhancement and investigation towards its practical use 
Sub Title (in English)  
Keyword(1) Single channel speech enhancement  
Keyword(2) Deep Neural Network  
Keyword(3) Unsupervised learning  
Keyword(4) Behavior analysis  
Keyword(5)  
Keyword(6)  
Keyword(7)  
Keyword(8)  
1st Author's Name Takuya Fujimura  
1st Author's Affiliation Nagoya University (Nagoya Univ.)
2nd Author's Name Tomoki Toda  
2nd Author's Affiliation Nagoya University (Nagoya Univ.)
3rd Author's Name  
3rd Author's Affiliation ()
4th Author's Name  
4th Author's Affiliation ()
5th Author's Name  
5th Author's Affiliation ()
6th Author's Name  
6th Author's Affiliation ()
7th Author's Name  
7th Author's Affiliation ()
8th Author's Name  
8th Author's Affiliation ()
9th Author's Name  
9th Author's Affiliation ()
10th Author's Name  
10th Author's Affiliation ()
11th Author's Name  
11th Author's Affiliation ()
12th Author's Name  
12th Author's Affiliation ()
13th Author's Name  
13th Author's Affiliation ()
14th Author's Name  
14th Author's Affiliation ()
15th Author's Name  
15th Author's Affiliation ()
16th Author's Name  
16th Author's Affiliation ()
17th Author's Name  
17th Author's Affiliation ()
18th Author's Name  
18th Author's Affiliation ()
19th Author's Name  
19th Author's Affiliation ()
20th Author's Name  
20th Author's Affiliation ()
Speaker Author-1 
Date Time 2023-03-01 11:00:00 
Presentation Time 20 minutes 
Registration for EA 
Paper # EA2022-112, SIP2022-156, SP2022-76 
Volume (vol) vol.122 
Number (no) no.387(EA), no.388(SIP), no.389(SP) 
Page pp.221-226 
#Pages
Date of Issue 2023-02-21 (EA, SIP, SP) 


[Return to Top Page]

[Return to IEICE Web Page]


The Institute of Electronics, Information and Communication Engineers (IEICE), Japan