Information theoretic similarity measures for content based image retrieval

Content-based image retrieval is based on the idea of extracting visual fea tures from image and using them to index images in a database. The comparis ons that determine similarity between images depend on the representations of the features and the definition of appropriate distance function. Most o f the research literature uses vectors as the predominate representation gi ven the rich theory of vector spaces. While vectors are an extremely useful representation, their use in large databases may be prohibitive given thei r usually large dimensions and similarity functions. In this paper, we prop ose similarity measures and an indexing algorithm based on information theo ry that permits an image to be represented as a single number. When use in conjunction with vectors, our method displays improved efficiency when quer ying large databases.