ELSNET-list archive

Category:   E-Conf
Subject:   Workshop on automatic transcription of speech
From:   euromap BE/NL
Email:   euromap_(on)_ntu.nl
Date received:   15 Mar 2002
Start date:   12 Apr 2002

-------------------------------------------------------------------- 1. WORKSHOP AUTOMATISCHE TRANSCRIPTIE VAN SPRAAK IN LEUVEN ********************************************************** ATRANOS workshop 12 april 2002 Faculty Club Leuven This workshop on the topic of automatic transcription of speech takes place at the University of Leuven and is sponsored by the AtraNoS project and CLIF (Computational Linguistics in Flanders). Programme: 10:30u Welcome: Patrick Wambacq (K.U.Leuven) 10:35u Introduction: Dirk Van Compernolle (K.U.Leuven) 10:45u The ALERT project: Gerhard Rigoll (Univ. Duisburg) 11:35u The DRUID project: Arjan Van Hessen (Univ. Twente) 12:25u Introduction to the ATRANOS project: Patrick Wambacq (K.U.Leuven) 12:45u Lunch 14:00u ATRANOS in depth contributions (20 mins per project partner) 15:20u Panel discussion 15:45u End, reception Participation is free, but registration is necessary (before 2nd of April). The number of participants is limited to 50. Please register by email to Annitta.DeMessem aeker@esat.kuleuven.ac.be and include name, affiliation and email address. Abstracts: The ALERT System: Audiovisual Broadcast Speech Transcription for Selective Dissemination of Multimedia Information This presentation gives a brief description of the ALERT system, which is under development by a consortium working on a research project sponsored by the European Commission. The ALERT system uses advanced speech recognition technology and video processing techniques in order to process large broadcast speech archives and multimedia information resources, from broadcast stations, TV shows and Internet sources, for the purpose of extracting specific information from such databases and inform selected customers with specific interest profiles about its contents. It is one of the most ambitious projects currently carried out in the Human Language Technologies (HLT) area (see also http://alert.uni-duisburg.de). The talk outlines the objectives of the overall system, its basic system architecture and the scientific approach taken in order to realize the specified demonstrators. This concerns new approaches for improving the current state-of-the-art in broadcast speech recognition, including novel methods for audio-visual segmentation of multimedia data directly into topic segments. Furthermore, new approaches to topic identification using a feature-based approach will be presented. The ALERT project is an impressive example in order to demonstrate how advanced speech recognition technology and other related methods can be already applied usefully in some of today's most competitive and emerging business sectors - the media monitoring market. DRUID: Multimedia Indexing and Retrieval on the basis of Image Processing & Language and Speech Technology DRUID aims at the development of tools for the indexing and retrieval of multimedia content. The results will serve the use and exploitation of digital mulitmedia archives. The research will advance existing insights and deploy the potential synergy between technologies from various expertise domains: language & speech technology, image processing and database technology. In terms of the objects to be made available for retrieval, the project will cover continuous objects containing text and/or speech elements (e.g. video and audio fragments), as well as static objects, ranging from paper documents and textually annotated images, to web page. Both fundamental research, as well as prototyping, tools testing and usability studies are part of the work plan. DRUID is a project within the scientific program of the Telematica Instituut. This institute is co-funded by the Dutch government and a number of industrial partners with the aim to advance state of the art in the area of ICT. ATraNoS: Automatic Transcription and Normalisation of Speech The objective of the AtraNoS project is to contribute to the development of better products for the automatic verbatim transcription of speech, and for the conversion of these transcriptions to a form that is better adapted to the needs of the end-user. One application which will be studied as a case study is the generation of subtitles for the benefit of hearing-impaired people. The scientific objectives of the project can be summarised as follows: (1) Develop a methodology for the reliable segmentation and labeling of continuous audio-streams in homogeneous segments containing one type of signal (e.g. speech from one speaker, double-talk, music, background noise, etc.). (2) Develop new methods for reducing the deteriorating effects of out-of-vocabulary words encountered during the automatic recognition of unrestricted speech. (3) Develop new methods for reducing the deteriorating effects of disfluencies encountered in spontaneous speech. (4) Develop methods for the automatic conversion of verbatim transcriptions to subtitles. AtraNoS is a project with funding from the Flemish government and is jointly carried out by the Katholieke Universiteit Leuven (ESAT/PSI speech group and CCL Centre for Computational Linguistics), the University of Gent (ELIS speech lab) and the University of Antwerp (CNTS research center). Travel directions: the Faculty club is in the Groot Begijnhof of the city of Leuven. Cars are not allowed in this part of the city, but parking space is available nearby on the Tervuursevest. Coming from the E314 highway, take exit 15 Leuven, at the third traffic lights just before the bridge turn right. At the next crossing (before the traffic lights) make a U turn. The parking entrance is just behind the Q8 petrol station. With the train, stop at Leuven station and take the taxi or bus nr. 2 and step out at the Naamsepoort. From there walk to the Faculty Club (see map at http://www.fc lub.kuleuven.ac.be/info/wegwijzer.html). More maps can be found at http:// cwisdb.cc.kuleuven.ac.be/kaarten-bin/basiskaart.pl (click to zoom in).

[print/pda] [no frame] [navigation table] [navigation frame]     Page generated 14-02-2008 by Steven Krauwer Disclaimer / Contact ELSNET