Skip navigation
The 'Holy Grail' of computer vision

Narrator:       This is Science Today. Engineers at the University of California, Santa Cruz have developed an algorithm that enables automated, quick recognition of an image or video. This powerful new approach could change the way people surf the Web for photos and videos.

Milanfar:        To some extent, this problem of being able to search for images with a single image has been, in a way, the Holy Grail of this particular field of research.

Narrator:       Peyman Milanfar, a professor of electrical engineering at UC Santa Cruz, led the team that developed this software.

Milanfar:        I think it's fair to say that for a fair amount of time, people simple didn't believe it was doable. That there was just not enough information in a single image so that you cold reliably find other things that are comparable in a way that we're presenting. We are currently in the process of patenting these results, and over the next few years, I can imagine this technology also being spun out, licensed to somebody perhaps and various commercial products build on top of it.

Narrator:       For Science Today, I'm Larissa Branin.