Retrieving distant objects using tree indices- an approach to exploratory databases

Developing content-based retrieval techniques for multimedia objects is an important research issue. So far, much attention has been paid to similarity search algorithms and indices. However, other types of interaction are also important to meet the demands in exploratory systems such as digital museums and online shops. In this paper, as an approach to exploratory databases, we study the problem to retrieve a given number of distant objects in high-dimensional vector space. Distant objects serve as sample objects which are expected to inspire users potential interests. We give an algorithm to retrieve distant objects using tree indices. The result of experimental studies shows the effectiveness of our algorithm