Retrieval Interfaces for Video Databases

Andrew S. Gordon and Eric A. Domeshek

Rich multimedia databases are useless without the ability to access the right piece of information at the right time. We have been investigating the development of indexing and retrieval systems for databases of stock video. By using a simplified conceptual indexing framework, we have been able to focus our efforts on the creation of semantic networks that organize and catalog stock video indexes and on the development of end-user indexing and retrieval interfaces. In this paper, we focus on the interface issues. We compare two classes of information retrieval interfaces: Queo' and Search interfaces and Zoom and Browse interfaces. Previous researchers have developed knowledge-rich retrieval systems that follow the Query and Search approach. We argue that it is better to offer choices than to play guessing games. Accordingly, we present a version of a Zoom and Browse interface that we have developed for the retrieval of stock video clips based on scene content indexes.

This page is copyrighted by AAAI. All rights reserved. Your use of this site constitutes acceptance of all of AAAI's terms and conditions and privacy policy.