A mock-up of an edited sequence (Real), generated by selecting the position "War in Afghanistan - For" and the point of view "Propagandist - Create Clash"
Examples of not edited interviews
A mock-up of an edited sequence (Real), generated by selecting the position "War in Afghanistan - For" and the point of view "Propagandist - Create Clash"
Examples of not edited interviews
This page presents information about VOX POPULI, a system for automatic generation of Biased Video Sequences. In few words, VOX POPULI uses rhetorical annotations to generate video sequences. Our annotation schema encodes the verbal information contained in the audio channel, identifying the claims the interviewees make and the argumentation structures they use to make those claims.Based on this schema, we construct a semantic graph which is traversed by rhetoric-based strategies selecting video segments.
The demo presents a Web interface to generate a short video documentary, as described in the paper. Please report problems to us.
As an example, a prerendered video sequence is here. The Toulmin analysis of the first part of this video sequence (as described in the above-mentioned paper) is the following:
VJ Cultuur is the result of a one-year long research about the culture of VJs in the Netherlands and consists of video interviews with 12 Dutch VJs. From the start of this project it was decided to present the video material in an interactive way using Vox Populi.
Interview (in Dutch) for the Dutch radio program Noorderlicht over scientific developments, broadcasted on national radio. A local copy of the interview is here.
Interview on the Amsterdam radio Onda Italiana in Dutch, unfortunately it was not recorded.
Interview (local copy here) with Elke van Cassel, Cursor (weekly information magazine of the University of Eindhoven), nr. 13, 2006.
There has been a press release about Vox Populi which has appeared on the following sites:
Fact Sheet (English) from CWI, the research center I worked for.
AlphaGalileo (local copy here) (English), independent resource for European research news.
Eindhoven Technical University (local copy here) (Dutch), where I defended my PhD.
Frontrunner (local copy here) (Dutch), the weblog of the Financieele Dagblad, a Dutch national newspaper.
ScienceGuide (not existing anymore) (Dutch), an online journal focussing on scientific contributions to the social and economic developments.
Edusite (local copy here) (Dutch), a site about developments in the IT sector in the international higher education.
Presentation given for my PhD defence, 30 November 2006, Technical University of Eindhoven.
Presentation given at Philips Research, 31 May 2006, Eindhoven.
Presentation given at the TOKEN Symposium 2006, 31 March 2006, Leiden University - Faculty of Law.
Presentation given at British Telecom Research, 26 Jan. 2006, Adastral Park, Ipswich, England
Presentation given at the tiende Conferentie Informatiewetenschap Nederland, 19 Nov 2005, Antwerpen
Presentation given at the Sixteenth ACM Conference on Hypertext and Hypermedia , 6-9 Sept. 2005, Salzburg
Presentation given at the IEEE International Conference on Multimedia and Expo (ICME), July 2005
My Thesis: Vox Populi: generating video documentaries from semantically annotated media repositories, defended on the 30th of November, 2006 at the Technical University of Eindhoven.
Stefano Bocconi, Frank Nack, and Lynda Hardman.
Supporting
the Generation of Argument Structure within Video Sequences. In:
Proceedings of the sixteenth ACM
Conference on Hypertext and Hypermedia 2005, September
2005
Abstract:
"Browsing is a useful way of exploring annotated media repositories. Sets of links can be automatically created from the annotations associated with the media items in the repository. When there are also relationships among the annotations themselves, such as when the annotation terms are part of a thesaurus, these relations can also be used in the link generation process. Using structured annotations and a thesaurus for generating link sets has two advantages. The first is to evaluate the effectiveness of the terms in the thesaurus for classifying the media items in the repository. The second is to be able to control the links being generated by changing relationships within the thesaurus. The work is illustrated using video segments annotated with argument structures, but we show that the method used is independent of the media types and applicable to systems that use similar annotation structures and typed relations among them."
Stefano Bocconi, Frank Nack, and Lynda Hardman.
Vox
populi: a tool for automatically generating video documentaries.
In: Proceedings of the
sixteenth ACM conference on Hypertext and hypermedia 2005,Salzburg,
Austria, September 2005, DEMONSTRATION SESSION: Demos
Abstract:
"Vox Populi is a system that automatically generates video documentaries. Our application domain is video interviews about controversial topics. Via a Web interface the user selects one of the possible topics and a point of view she would like the generated sequence to present, and the engine selects and assembles video material from the repository to satisfy the user request."
Stefano Bocconi, Frank Nack, and Lynda Hardman.
Using
Rhetorical Annotations for Generating Video Documentaries. In:
Proceedings of the IEEE
International Conference on Multimedia and Expo (ICME) 2005,
Amsterdam, Netherlands, July 2005
Abstract:
"We use rhetorical annotations to specify a generation process that can assemble meaningful video sequences with a communicative goal and an argumentative progression. Our annotation schema encodes the verbal information contained in the audio channel, identifying the claims the interviewees make and the argumentation structures they use to make those claims. Based on this schema, we construct a semantic graph which is traversed by rhetoric-based strategies selecting video segments. The selected video segments are edited to form a meaningful video sequence."
Presentation given at the Media Fabrics group at Massachusetts Institute of Technology, Cambridge, Massachusetts, USA - October 2004.
Presentation given at the Fraunhofer Institute for Applied Information Technology FIT, Bonn, Germany - February 2004.
Stefano Bocconi. VOX POPULI: Automatic Generation of Biased Video Sequences. In: First ACM Workshop on Story Representation, Mechanism and Context (pages 9-16), October 2004 application/pdf Abstract:
"We describe our experimental rhetoric engine Vox Populi that generates biased video-sequences from a repository of video interviews and other related audio-visual web sources. Users are thus able to explore their own opinions on controversial topics covered by the repository. The repository contains interviews with United States residents stating their opinion on the events occurring after the terrorist attack on the United States on the 11th of September 2001. We present a model for biased documentary statements, such as interviews, and explain in detail how this model facilitates the automatic generation of rhetorical arguments on a micro-level. We outline the required representations of relevant rhetorical structures and the way they can be processed. The processes are described via examples generated by our experimental engine. The first example shows how to logically counter an opinion using semantics contained in the audio tracks from the database, while the second example describes the generation of an emotional counterargument using visual material."
The annotation are encoded in RDF; the schema is defined here and the instances here; we used Protégé 2000 to edit the RDF files (you can see a screenshot); we used Sesame to store and query our knowledge-base, you can browse and query the same knowledge-base we use here (try for ex. the Explore link). To select all the statements contained in the knoledge-base, try to cut and paste the following SeRQL query here:
select X, SubjectDescription, ModifierDescription, PredicateDescription from {X} VoxPopuli:subject {} VoxPopuli:partDescription {SubjectDescription}, [{X} VoxPopuli:modifier {Modifier} VoxPopuli:partDescription {ModifierDescription}], [{X} VoxPopuli:predicate {Predicate} VoxPopuli:partDescription {PredicateDescription}] using namespace MediaClipping = <http://www.w3.org/2001/SMIL20/MediaClipping#>, VoxPopuli = <http://www.cwi.nl/~media/ns/VP/VoxPopuli.rdfs#>, BasicMedia = <http://www.w3.org/2001/SMIL20/BasicMedia#>
