tineye1

TinEye: Tracking Your Images, Pixel-By-Pixel

A new image-based search engine can help you see who’s using your images — even if they’ve been cropped, resized, or otherwise altered.

TinEye uses “sophisticated pattern recognition algorithms” to find any instance of an image anywhere on the Internet. As its creators explain it, the service creates “a compact digital signature or ‘fingerprint,'” then compares “its fingerprint to the fingerprint of every single other image in the TinEye search index.”

A complex system, but put simply: You give it an image and it’ll show you who else has it.

Here’s what happens: You sign up for a free account and wait a few minutes for a confirmation e-mail. Then, you login, upload an image or input an image URL — and within seconds, TinEye will show you every place the same image is being used. Still in beta, the service is still somewhat limited, though already fairly impressive: The database is at just over 700 million images and is growing by the day as its spiders continue to crawl the Web for images.

I first tried a search for The Inquisitr logo, seen in the banner at the top of this site. TinEye says no one else is using it. Next, I tried one certain to deliver plenty of offenders: the Miley Cyrus Vanity Fair photo. Sure enough, TinEye found a few dozen sites that’d lifted the picture and published it themselves.

Some other hot searches TinEye points out include the Uncle Sam “I Want You” poster, the Mona Lisa, the U.S. dollar bill, and the Hello Kitty character.

Canada-based Idee is behind this new startup. The company focuses on image identification software and has created a few other products utilizing similar concepts. Its founders say their technology can handle low-resolution or grainy images and still be able to recognize them.


Comments