The University of Chicago Magazine
Want to find an image on the World Wide Web? Try "WebSeer," an image-based search engine designed and built at the U of C's Intelligent Information Laboratory. Michael Swain, assistant professor in computer science and the main architect of WebSeer, calls it "the first useful way that people can search for images: photographs as well as drawings."
WebSeer <http://webseer.cs.uchicago.edu> stores contextual information about images--the caption, Web-page title, and image-file name. It also gives information about the image--whether it's a photo or a drawing, color or black-and-white, whether it includes people's faces, and if so, how many and how close they are in the frame.
"If you just look at the pixels of an image, it's beyond the current computer vision technology to identify particular objects such as elephants or Saabs or Zeppelins," says third-year
WebSeer's sophisticated automatic indexing system can add tens of thousands of images a day to its database, which now includes 500,000 images.
Here's how it works. A search for, say, the aurora borealis, or northern lights, yields several photographs--plus a diagram of how the lights are generated--and links to pages on the Space Plasma Physics Center, tourism in Scandinavia, and NASA studies of solar wind and auroral physics. Clicking on the image's thumbnail takes you directly to the image site.
Already, researchers from Berkeley, MIT, Caltech, and Carnegie Mellon, as well as the U of C, are plugging applications into the system. Predicts Michael Swain: "WebSeer is going to be the most widely used application of computer vision ever."
In this department:
Also in the Journal:
Return to February 1997 Table of Contents