FIRE – flexible image retrieval engine: ImageCLEF 2004 evaluation

  • Authors:
  • Thomas Deselaers;Daniel Keysers;Hermann Ney

  • Affiliations:
  • Lehrstuhl für Informatik VI – Computer Science Department, RWTH Aachen University, Aachen, Germany;Lehrstuhl für Informatik VI – Computer Science Department, RWTH Aachen University, Aachen, Germany;Lehrstuhl für Informatik VI – Computer Science Department, RWTH Aachen University, Aachen, Germany

  • Venue:
  • CLEF'04 Proceedings of the 5th conference on Cross-Language Evaluation Forum: multilingual Information Access for Text, Speech and Images
  • Year:
  • 2004

Quantified Score

Hi-index 0.00

Visualization

Abstract

We describe FIRE, a content-based image retrieval system, and the methods we used within this system in the ImageCLEF 2004 evaluation. In FIRE, various features are available to represent images. The diversity of available features allows the user to adapt the system to the task at hand. A weighted combination of features admits flexible query formulations and helps with processing specific queries. For the ImageCLEF 2004 evaluation, we used the image content alone and obtained the best result in the category “only visual features, fully automatic retrieval” in the medical retrieval task. Additionally, the results compare favorably to other systems, even if they make use of the textual information in addition to the images.