Home Contents Search

Access
Up ] QoS ] Scaling/Security ] [ Access ] Computation ]

 

Proposal
Current Status

Structured Access to Information

The third area of fundamental importance for the software infrastructure is the organization of textual and image data based on structural properties. Three research groups are investigating questions in this area. Seshadri is building next-generation search engines that operate on structured heterogeneous databases, including image, multimedia and geometric data. Zabih is investigating techniques for searching large repositories of images based on properties such as color, texture and shape. Huttenlocher is applying visual matching techniques to interactive applications involving images and video, and is developing tools to support group collaboration. Huttenlocher and Smith plan to provide video browsers that enable reference to and search of oral presentations based on their content. Determining the applicability of such new information resources requires large numbers of users who place serious demands on the systems, such as students reviewing material for an exam.

Participants

Praveen Seshadri, Assistant Professor, Computer Science
Ramin Zabih, Assistant Professor, Computer Science
Daniel Huttenlocher, Associate Professor, Computer Science

Web Links

PREDATOR object-relational DBMS
Computer Vision Group at Cornell University
 

 

Back Home Up Next

Last modified on: 07/30/99