Extraction of Association Rules between Low-Level Descriptors and Semantic Descriptors in an Image Database

In this paper we combine previous works are done in the domain of image processing used to extract automatically low-level descriptors and previous works in the eld of data mining used to nd a kind of dependencies between variables al led association rules. Using our own prototypes we extract low-level descriptors, and then extract association rules between these descriptors and semantic descriptors. We run experiments on an image database containing 532 paintings. These experiments show that the combination of techniques proposed in this paper an be used to extract eÆiently dependencies between low-level descriptors (colour descriptors) and semantic descriptors (names of painters)