Chapter

Machine Learning for Multimodal Interaction

Volume 5237 of the series Lecture Notes in Computer Science pp 260-271

A Probabilistic Model for User Relevance Feedback on Image Retrieval

  • Roberto ParedesAffiliated withPattern Recognition and Human Language Technology Group, Universidad Politécnica de Valencia
  • , Thomas DeselaersAffiliated withHuman Language Technology and Pattern Recognition Group Computer Science Department, RWTH Aachen University
  • , Enrique VidalAffiliated withPattern Recognition and Human Language Technology Group, Universidad Politécnica de Valencia

Abstract

We present a novel probabilistic model for user interaction in image retrieval applications which accounts for consistency among the retrieved images and considers the distribution of images in the database which is searched for. Common models for relevance feedback do not consider this and thus do not incorporate all available information. The proposed method is evaluated on two publicly available benchmark databases and clearly outperforms recent competitive methods.