Faculty of Information Technology, BUT

Publication Details

Transcribing Meetings with the AMIDA System

HAIN Thomas, BURGET Lukáš, DINES John, GARNER Phillip N., GRÉZL František, EL Hannani Asmaa, HUIJBREGTS Marijn, KARAFIÁT Martin, LINCOLN Mike and WAN Vincent. Transcribing Meetings with the AMIDA System. IEEE Transactions on Audio, Speech, and Language Processing, vol. 20, no. 2, pp. 486-498. ISSN 1558-7916. Available from: http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=5983475
Czech title
Rozpoznávání meetingů se systémy AMIDA
Type
journal article
Language
english
Authors
Hain Thomas (USF)
Burget Lukáš, doc. Ing., Ph.D. (DCGM FIT BUT)
Dines John (IDIAP)
Garner Phillip N. (IDIAP)
Grézl František, Ing., Ph.D. (DCGM FIT BUT)
El Hannani Asmaa (USF)
Huijbregts Marijn (UTWENTE)
Karafiát Martin, Ing., Ph.D. (DCGM FIT BUT)
Lincoln Mike (IDIAP)
Wan Vincent (USF)
URL
Keywords
AMI corpus, Juicer, meeting transcription, multiple distant microphone, resource optimisation, rich text
Abstract
This paper describes AMIDA systems for transcription of conference and lecture room meetings that were developed for participation in the RT evaluations conducted by NIST in years 2007 and 2009.

Annotation
In this paper, we give an overview of the AMIDA systems for transcription of conference and lecture room meetings. The systems were developed for participation in the Rich Transcription evaluations conducted by the National Institute for Standards and Technology in the years 2007 and 2009 and can process close talking and far field microphone recordings. The paper first discusses fundamental properties of meeting data with special focus on the AMI/AMIDA corpora. This is followed by a description and analysis of improved processing and modeling, with focus on techniques specifically addressing meeting transcription issues such as multi-room recordings or domain variability. In 2007 and 2009, two different strategies of systems building were followed. While in 2007 we used our traditional style system design based on cross adaptation, the 2009 systems were constructed semi-automatically, supported by improved decoders and a new method for system representation. Overall these changes gave a 6%-13% relative reduction in word error rate compared to our 2007 results while at the same time requiring less training material and reducing the real-time factor by five times. The meeting transcription systems are available at www.webasr.org.
Published
2012
Pages
486-498
Journal
IEEE Transactions on Audio, Speech, and Language Processing, vol. 20, no. 2, ISSN 1558-7916
Book
IEEE Transactions on Audio, Speech, and Language Processing
Publisher
IEEE Signal Processing Society
Place
New York, US
DOI
BibTeX
@ARTICLE{FITPUB9945,
   author = "Thomas Hain and Luk\'{a}\v{s} Burget and John Dines and N. Phillip Garner and Franti\v{s}ek Gr\'{e}zl and Asmaa Hannani El and Marijn Huijbregts and Martin Karafi\'{a}t and Mike Lincoln and Vincent Wan",
   title = "Transcribing Meetings with the AMIDA System",
   pages = "486--498",
   booktitle = "IEEE Transactions on Audio, Speech, and Language Processing",
   journal = "IEEE Transactions on Audio, Speech, and Language Processing",
   volume = 20,
   number = 2,
   year = 2012,
   location = "New York, US",
   publisher = "IEEE Signal Processing Society",
   ISSN = "1558-7916",
   doi = "10.1109/TASL.2011.2163395",
   language = "english",
   url = "https://www.fit.vut.cz/research/publication/9945"
}
Back to top