Categorical Image Search


Shawn Newsam, Baris Sumengen, B.S.Manjunath


This project investigated multi-modal and multi-feature access in images collected from the WWW. Relevance feedback was used to iteratively improve the results online.

This system was designed and implemented by Baris Sumengen and Shawn Newsam.


» Demo (include 607,000 images)

With the proliferation of digital images and video, there is a real need for tools that help navigate such large volumes of data. The emphasis we have taken in our work is to provide such access using localized image/video information (see the related work on segmentation). The image features of interest include the dominant color in the regions of interest and the region texture. Both these descriptors have been accepted to the MPEG-7 experimental Model (XM).


This work is supported in part by an ONR/AASERT #N00014-98-1-0515 grant.