Information: Join today and make your research activities more affordable! Technical workshop participation fees and annual registration fees are available at member rates.
Notice: [Important] Announcement of Changes to Registration Fee Payment and Manuscript Upload Procedures for IEICE Technical Meetings
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 2008-03-20 15:15
[Poster Presentation] LVCSR based on Context-Dependent Syllable Acoustic Models
Jian Zhang, Longbiao Wang, Seiichi Nakagawa (Toyohashi Univ. of Tech.) SP2007-200
Abstract (in Japanese) (See Japanese page) 
(in English) We propose an effective and accurate inter-word context dependent modeling for large vocabulary continuous speech recognition (LVCSR). As well known, intra-word context dependent modeling can be realized by describing the context dependent syllables in the dictionary. However, it usually suffers from the limitations of less accuracy because it does not model inter-syllable pronunciation variations. In our previous study, a combinational use of linear lexicon and tree-structured lexicon in a 1-best approximation search algorithm for LVCSR was proposed. We only need to make branches for the head syllable according to the contexts and the paths are merged at the second syllable for the linear lexicon. For the tree-structured lexicon, branches are made in a similar way. At the end node of a word the language scores have to be compensated considering the inter-word context, but the scores of contexts other than that of the best history have lost because of the merge at the second syllable. To solve this problem, we introduce the `likelihood difference index'. We also investigate the effect of rescoring of the context dependent syllable acoustic model in the 2nd pass mode. The proposed algorithms were evaluated on JNAS and CSJ corpus. The proposed algorithms obtained a remarkable improvement of recognition performance, and the rescoring of the context dependent syllable acoustic models in the 2nd pass mode also achieved a further improvement even the same acoustic models were used in the 1st pass.
Keyword (in Japanese) (See Japanese page) 
(in English) LVCSR / context-dependent acoustic models / acoustic and language model rescoring / linear lexicon / tree-structured lexicon / / /  
Reference Info. IEICE Tech. Rep., vol. 107, no. 551, SP2007-200, pp. 81-86, March 2008.
Paper # SP2007-200 
Date of Issue 2008-03-13 (SP) 
ISSN Print edition: ISSN 0913-5685    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 SP2007-200

Conference Information
Committee SP  
Conference Date 2008-03-20 - 2008-03-21 
Place (in Japanese) (See Japanese page) 
Place (in English) Univ. Tokyo 
Topics (in Japanese) (See Japanese page) 
Topics (in English) International Workshop (Mar 20), Speech Production, Speech Perception, Hearing and Speech, etc. (Mar 21) 
Paper Information
Registration To SP 
Conference Code 2008-03-SP 
Language English 
Title (in Japanese) (See Japanese page) 
Sub Title (in Japanese) (See Japanese page) 
Title (in English) LVCSR based on Context-Dependent Syllable Acoustic Models 
Sub Title (in English)  
Keyword(1) LVCSR  
Keyword(2) context-dependent acoustic models  
Keyword(3) acoustic and language model rescoring  
Keyword(4) linear lexicon  
Keyword(5) tree-structured lexicon  
Keyword(6)  
Keyword(7)  
Keyword(8)  
1st Author's Name Jian Zhang  
1st Author's Affiliation Toyohashi University of Technology (Toyohashi Univ. of Tech.)
2nd Author's Name Longbiao Wang  
2nd Author's Affiliation Toyohashi University of Technology (Toyohashi Univ. of Tech.)
3rd Author's Name Seiichi Nakagawa  
3rd Author's Affiliation Toyohashi University of Technology (Toyohashi Univ. of Tech.)
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 ()
21st Author's Name  
21st Author's Affiliation ()
22nd Author's Name  
22nd Author's Affiliation ()
23rd Author's Name  
23rd Author's Affiliation ()
24th Author's Name  
24th Author's Affiliation ()
25th Author's Name  
25th Author's Affiliation ()
26th Author's Name / /
26th Author's Affiliation ()
()
27th Author's Name / /
27th Author's Affiliation ()
()
28th Author's Name / /
28th Author's Affiliation ()
()
29th Author's Name / /
29th Author's Affiliation ()
()
30th Author's Name / /
30th Author's Affiliation ()
()
31st Author's Name / /
31st Author's Affiliation ()
()
32nd Author's Name / /
32nd Author's Affiliation ()
()
33rd Author's Name / /
33rd Author's Affiliation ()
()
34th Author's Name / /
34th Author's Affiliation ()
()
35th Author's Name / /
35th Author's Affiliation ()
()
36th Author's Name / /
36th Author's Affiliation ()
()
Speaker Author-1 
Date Time 2008-03-20 15:15:00 
Presentation Time 90 minutes 
Registration for SP 
Paper # SP2007-200 
Volume (vol) vol.107 
Number (no) no.551 
Page pp.81-86 
#Pages 6 
Date of Issue 2008-03-13 (SP) 

[Return to Top Page]

[Return to IEICE Web Page]
The Institute of Electronics, Information and Communication Engineers (IEICE), Japan