Repository hosted by TU Delft Library

Home · Contact · About · Disclaimer ·
 

Multimedia content with a speech track: ACM multimedia 2010 workshop on searching spontaneous conversational speech

Publication files not online:

Author: Larson, M. · Ordelman, R. · Metze, F. · Kraaij, W. · Jong, F. de
Type:article
Date:2010
Institution: TNO Informatie- en Communicatietechnologie
Source:18th ACM International Conference on Multimedia ACM Multimedia 2010, MM'10, 25-29 October 2010, Firenze, Italy, 1747-1748
Identifier: 426518
doi: DOI:10.1145/1873951.1874351
Keywords: Informatics · Audio-visual retrieval · Multimedia access · Speech · Speech recognition · Spoken content

Abstract

When multimedia content has a speech track, a whole array of techniques involving speech recognition and analysis becomes available for indexing and structuring and can provide users with improved access and search. The set of new domains standing to benefit from these techniques encompasses talkshows, lectures, meetings, interviews, debates, podcasts, call center recordings, cultural heritage archives, social video on theWeb and spoken natural language queries. The success that has been achieved in the area of broadcast news retrieval [3] has yet to be extended to providing users with access and retrieval in all domains. New domains are challenging because they often involve unplanned, spontaneously produced speech and/or a conversational setting. The goal of the SSCS workshop series on Searching Spontaneous Conversational Speech is to help push forward research on these new domains by promoting cross-pollination between researchers from the areas of speech recognition, audio processing, multimedia analysis and information retrieval. After the success of the initial SSCS workshop held at SIGIR 2007 [2], a second workshop followed at SIGIR 2008 [4]. In order to actively connect with the multimedia community, SSCS 2009 [5] was held at ACM Multimedia