‘Quick, who painted fish?’: searching a picture database with the QBIC project at UC Davis

The University of California ar Davis's Department of Art and Art History is using QBIC (Query by Image Content), a visual query software program currently under development by IBM, to determine how effective it may prove in retrieving art images based on what they look like, rather than relying on text indexing. Beginning in January, 1992, in a joint research project with IBM, we began the construction of a pilot database of art images selected from our slide library. This paper will describe the construction of our pilot database, and will present preliminary findings