Utilisation des treillis de Galois pour la caractérisation d’ensembles d’images Characterization of Image Sets : The Galois Lattice Approach

This paper presents a new method for supervised image classification. To each image set is attached a landmark with the aim to characterize and to distinguish it from each others. The different properties, deduced from image primitives and their relationships with the sets of images are structured and organized into a hierarchy thanks to an original method relying on a mathematical formalism called Galois (or Concept) Lattices. Such lattices allow us to select properties as landmarks of specific classes. This approach belongs to the mobile robotic field and topological localization in a structured environment, for which each set of images is issued from one place that must be characterized. This paper exposes this original process and results after experimentation.