Combining SVM and graph matching in a Bayesian multiple classifier system for image content recognition

Abstract

In this paper, we propose an approach to image content recognition that exploits the benefits of different image representations to associate meaning with images. We choose classifiers based on global appearance, scene structure and region type occurrence, and define confidence measures on their output. The resulting posterior probabilities of the classifiers are combined in a Bayesian framework. We show that this method leads to a robust and efficient system that contributes to reducing the semantic gap between low level image features and higher level image descriptions. © Springer-Verlag Berlin Heidelberg 2006.