Spatio-temporal decomposition of sport events for video indexing

  • Authors:
  • Lluis Barceló;Xavier Orriols;Xavier Binefa

  • Affiliations:
  • Unitat de Processament d'Imatges i Intel.ligència Artificial and Dept. d'Informàtica, Edifici Q, Universitat Autònoma de Barcelona, Bellaterra, Barcelona, Spain;Unitat de Processament d'Imatges i Intel.ligència Artificial and Dept. d'Informàtica, Edifici Q, Universitat Autònoma de Barcelona, Bellaterra, Barcelona, Spain;Unitat de Processament d'Imatges i Intel.ligència Artificial and Dept. d'Informàtica, Edifici Q, Universitat Autònoma de Barcelona, Bellaterra, Barcelona, Spain

  • Venue:
  • CIVR'03 Proceedings of the 2nd international conference on Image and video retrieval
  • Year:
  • 2003

Quantified Score

Hi-index 0.00

Visualization

Abstract

In this paper we present a robust technique for summarizing sport video sequences. Unlike dense optical flow and parametric methods, we develop a semi-automatic application that uses geometrical information, such as straight lines, extracted from sequences of images. This information is used to compute the homographies between consecutive frames. This estimation yields a manner of synthesizing a high resolution image of the background plus the field-centered trajectories of the moving objects onto it.