Selecting good keys for triangle-inequality-based pruning algorithms

A new class of algorithms based on triangle inequality has been proposed for use in content-based image retrieval. These algorithms rely on comparing a set of key images to the database images, and storing the computed distances. Query images are later compared to the keys, and the triangle inequality is used to speedily compute lower bounds on the distance from the query to each of the database images. This paper addresses the question of increasing performance of this algorithm by the selection of appropriate key images. Several algorithms for key selection are proposed and tested.