ImageCLEF 2009 Robot Vision Task - Approaches

Information about the approaches used by the participants can be found in the CLEF'09 Working Notes:

• University of Glasgow at Image CLEF 2009 Robot Vision Task
  Yue Feng, Martin Halvey and Joemon M. Jose
• UAIC: Participation in ImageCLEF 2009 Robot Vision Task
  Emanuela Boroş, George Roşca and Adrian Iftene
• Fast LSIS Profile Entropy Features for Robot Visual Self-Localization
  Herve Glotin, Zhong-Qiu Zhao and Emilie Dumonti
• A Particle-Filter Based Self-Localization Method using Invariant Features as Visual Information
  Jesús Martínez-Gómez, Alejando Jiménez-Picazo, and Ismael García-Varea
• Visual Language Modeling for Mobile Localization: LIG Participation in RobotVision’09
  Trong-Ton Pham, Loïc Maisonnasse and Philippe Mulhem

Several multimedia extensions visualizing the approaches and the results obtained by the participating teams can be found below.

Idiap Research Institute, Martigny, Switzerland

The video below presents the recognition results for one of the experiments performed on the IDOL2 database. The approach presented here is based on a single global visual cue (Composed Receptive Field Histograms) and models of places built using Support Vector Machines.


Please see the following publication for more details about the approach:
A. Pronobis, O. Martínez Mozos, and B. Caputo. SVM-based discriminative accumulation scheme for place recognition. In Proceedings of the IEEE International Conference on Robotics and Automation (ICRA08), Pasadena, CA, USA, May 2008.