Sketch-based image queries in topographic databases

In this paper we present the development of a system prototype for sketch-b ased queries for the content-based retrieval of digital images from topogra phic databases. We discuss our overall strategy and associated algorithmic and implementation aspects, and we present associated database design issue s. The query tools devised in this research are employing user-provided ske tches of the shape and spatial configuration of the object(s) which should appear in the images to be retrieved. Our matching tool is inspired by leas t-squares matching (Ism) and represents an extension of Ism to function wit h a variety of raster representations. Our strategy makes use of a hierarch ical organization of feature shapes within a feature library. The results a re ranked according to statistical scores and the user can subsequently nar row or broaden his/her search according to the previously obtained results and the purpose of the search. Our approach combines the design of an integ rated database environment with the development of a feature library and th e necessary matching tools. We discuss our overall strategy and individual database components and present some implementation results. (C) 1999 Acade mic Press.