Queries by visual content using fuzzy similarity measures on regional descriptions
Résumé
This paper introduces a new CBIR system based on segmented images and fuzzy similarity measures. It
gives the possibility to the user to specify a set of regions, and to retrieve images containing these regions in an
approximate spatial configuration. It also supports competitive queries, and allows the expert-user to manipulate the
parameters of the functions involved. In this manuscript we focus in the architectural aspects of the system.