Interactive Indexing Into Image Databases

The general problem of object recognition is diicult and often requires a large amount of computing resources, even for locating an object within a single image. How, then, can it be possible to build a tool to for indexing into a a large database of, say, thousand of images, which works eeectively in \interactive time" on aaordable hardware? One important optimization is to take advantage of interaction with the user to nd out what types of variation are expected in the database, and to rely on the user to discriminate between similar-looking objects. Another is to create appropriate data structures oo-line to speed on-line searches. We are building a tool, called FINDIT, for locating the image of an object from within a large number of images of scenes which may contain the object. The user outlines and object in an image that he wants to nd in the database, and speciies the constraints on the transformations of the object that are expected to occur. The program acts as a lter to quickly reduce the possible number of candidates to a number small enough to be perused by the user. FINDIT will choose an appropriate search algorithm depending on the selection of constraints by the user.