Tuesday, February 1, 2011

TinEye- Image search for exploring attribution?

I wrote about using Google Goggles to find the source of the image that we used for Capstone visuals. ITeam members were impressed when I showed them how my phone could find the image. But they typically don't have smartphones that are capable of running Goggles.

In reviewing Nevit's contributions to WikiMedia Commons, I learned about TinEye.com, a browser-based reverse search engine. I want ITeam members to get some experience with powerful applications of technologies that will help with a problem they face in school. So, I tested TinEye on the same photos that Goggles found. When TinEye could not find the modified image (arrow added to point to the keystone), I cropped back to the largest part of the unmodified image.

An Awesome screenshot annotated to show the contrast with Goggles results.

Contrast it with the original:

Stone Arch Bridge in Yuanmingyuan

By Shizhao (Own work) [CC-BY-SA-2.5], via Wikimedia Commons

If TinEye can survive the competition with Google Goggles, then I suspect it will be able to locate photos like this. While we wait for the dust to settle, TinEye provides some Cool Searches that illustrate the power of artificial intelligence for current and future challenges.

No comments: