This paper describes how we maintain color and spatial index information on more than 1,000,000 images and how we allow users to browse the spatial color feature space. We break down all our images in color-based quad trees and we store all quad trees in our main-memory database. We allow users to browse the quad trees directly, or they can pre-select images through our color bit vector, which acts as an index accelerator. A Java based textsc{gui is used to navigate through our image indexes.

, ,
Information Systems [INS]
Database Architectures

Bosch, H.G.P, Nes, N.J, & Kersten, M.L. (2000). Navigating through a forest of quad trees to spot images in a database. Information Systems [INS]. CWI.