Summer 1998 Research 

So you want to know what I'm up to this summer?

I'm spending this summer at the University of Iowa in Iowa City, IA  as part of the Computer Science Research Experience for Undergraduates.  My research is in the department of Library and Information Sciences at UI.  I am working with Dr. David Eichmann and Waka Waktola to further develop Spider, a visual information retrieval system for the World Wide Web.  Our project has been named VIRM (Visual Information Retrieval Machine)

 

How does VIRM work?

VIRM is a Java-based metasearch engine that takes a keyword and purges HTML documents for image URLs (the .GIF, .JPG, and .JPEG files).  These images are then clustered according to color similarity so that the user can search through the images with relative ease.  So, for example, if a user wanted to find images of lions, she or he would enter the word "lion" into the keyword field.  Spider will then call other search engines to find the 10 - 100 top document matches for the word "lion."  The URLs of these documents are returned to Spider, which sifts through the document and picks out the image URLs, which are then grouped by similarity.  The reason for this is that many of the pictures on a page about lions may not actually be pictures of lions -- they may be buttons, lines,  a zoo logo, etc.  By clustering the images, the user will be able to point the crawler towards the images that are actually lions.

 For more information about this, read our report on the Visual Information Retrieval Machine (VIRM): Content-Based Image Retrieval on the World Wide Web.

 

Where can I see some examples of visual information retrieval (VIR)?

Here are some other places that have done research on VIR and VIM (Visual Information Management).  Check it out!

   

 

---