Warning: your browser seems to be quite old, or at least old enough that it does not support the minimal standard capabilities of today's browsers, which are used extensively on this site. Please upgrade to a browser that supports web standards, specifically one that supports minimal CSS1 capabilities, such as Netscape 6 or higher, IE 5 or higher, or Opera 5 or higher. Your browsing experience will be much better! Thanks.

Title Automatic paradigms for semantic annotation.

Starting date October 2005.

Ending date April 2008.


Main activities
WG4 carries out a collaborative research on automatic paradigms for semantic annotation, focusing on:


The main activity within WG4 has been the participation to the TRECVID video retrieval evaluation organized by the U.S. National Institute of Standards and Technology. Our first participation involved three separate systems from Bilkent University, National Technical University of Athens (NTUA), and Middle East Technical University (METU) for classifying shots into high-level classes based on their content. These systems involved technologies such as image segmentation, clustering, latent semantic analysis, text localization, optical character recognition, Bayesian and neural network classifiers.

Initial work targeted high-level concepts such as snow, vegetation, waterscape, sky, mountain, outdoor, desert, road, fire, explosion, government leader and corporate leader. Results from three systems were independently submitted to the TRECVID evaluation.

Future work will concentrate on the fusion of different advantages of such systems for a unified approach. Partners that are working together toward that goal are Bilkent University, Queen Mary University of London, National Technical University of Athens, Delft University of Technology, Universidade da Beira Interior, University of Zilina, Telefonica I+D, University of Novi Sad, University of Belgrade, and Fondazione Ugo Bordoni. We are targeting all 37 concepts included in TRECVID 2007: Sports, Entertainment, Weather, Court, Office, Meeting, Studio, Outdoor, Building, Desert, Vegetation, Mountain, Road, Sky, Snow, Urban, Waterscape and Waterfront, Crowd, Face, Person, Police and Security, Military, Prisoner, Animal, Computer and TV screen, US Flag, Airplane, Car, Bus, Truck, Boat and Ship, Walking and Running, People Marching, Explosion and Fire, Natural Disaster, Maps, Charts.


Results of our joint work for TRECVID 2006 are given in our summary paper:
Download TRECVID 2006 summary paper

Mailing list for WG4