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 2021-03-05 09:45
Improved Speech Separation Performance from Monaural Mixed Speech Based on Deep Embedding Network
Shaoxiang Dang, Tetsuya Matsumoto, Hiroaki Kudo (Nagoya Univ.), Yoshinori Takeuchi (Daido Univ.) PRMU2020-85
Abstract (in Japanese) (See Japanese page) 
(in English) Speech separation refers to the separation of utterances in which multiple people are speaking simultaneously. The idea of deep clustering (DC) is put forward by using a deep embedding network to embed audio data in the underlying manifold, and data with similar property gathers tightly in the embedding space. Then this model uses a clustering algorithm because the clustering algorithm can easily separate distributed data. Regarding the learning process, the model is supervised by an ideal affinity matrix constructed of binary masks of annotation data. However, the binary mask gives a bottleneck to the entire system since the same position of bins in masks are assigned to 0 or 1 according to the contribution of individual utterances to mixed spectrograms. Thus, we propose an extended two-stage version of network based on the deep embedding. The network can eliminate the shortcomings by using various more accurate masks. We employ DC as our first stage, and conduct a permutation invariant training approach to prevent permutation problem in the second stage. As a result, the results according to our experiment outperforms the original DC model by 1.55dB in SNR by 4.45dB in SDR, 4.41dB in SDRi, 0.16 in STOI, and 0.3 in PESQ on average. We also observe that the proposed method can recover the defects in the spectrograms brought in by DC.
Keyword (in Japanese) (See Japanese page) 
(in English) speech separation / deep embedding network / monaural speech separation / permutation invariant training / / / /  
Reference Info. IEICE Tech. Rep., vol. 120, no. 409, PRMU2020-85, pp. 91-96, March 2021.
Paper # PRMU2020-85 
Date of Issue 2021-02-25 (PRMU) 
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 PRMU2020-85

Conference Information
Committee PRMU IPSJ-CVIM  
Conference Date 2021-03-04 - 2021-03-05 
Place (in Japanese) (See Japanese page) 
Place (in English) Online 
Topics (in Japanese) (See Japanese page) 
Topics (in English) Computer Vision and Pattern Recognition for specific environment 
Paper Information
Registration To PRMU 
Conference Code 2021-03-PRMU-CVIM 
Language English 
Title (in Japanese) (See Japanese page) 
Sub Title (in Japanese) (See Japanese page) 
Title (in English) Improved Speech Separation Performance from Monaural Mixed Speech Based on Deep Embedding Network 
Sub Title (in English)  
Keyword(1) speech separation  
Keyword(2) deep embedding network  
Keyword(3) monaural speech separation  
Keyword(4) permutation invariant training  
Keyword(5)  
Keyword(6)  
Keyword(7)  
Keyword(8)  
1st Author's Name Shaoxiang Dang  
1st Author's Affiliation Nagoya University (Nagoya Univ.)
2nd Author's Name Tetsuya Matsumoto  
2nd Author's Affiliation Nagoya University (Nagoya Univ.)
3rd Author's Name Hiroaki Kudo  
3rd Author's Affiliation Nagoya University (Nagoya Univ.)
4th Author's Name Yoshinori Takeuchi  
4th Author's Affiliation Daido University (Daido Univ.)
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 2021-03-05 09:45:00 
Presentation Time 15 minutes 
Registration for PRMU 
Paper # PRMU2020-85 
Volume (vol) vol.120 
Number (no) no.409 
Page pp.91-96 
#Pages
Date of Issue 2021-02-25 (PRMU) 


[Return to Top Page]

[Return to IEICE Web Page]


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