Idée inc. has launched the private beta of TinEye, an Image identification search engine built upon Amazon's AWS platform.

Rather than relying on tags and keywords, TinEye search takes a user-supplied image and searches for all the places where that same image has been used, regardless of changes made to the actual file. If the photo has been cropped, resized, re-colored, or altered, TinEye still discovers it.

Powered by Amazon's Elastic Compute Cloud and Simple Storage Service, the engine searches through nearly 500 million images in a matter of seconds.

High-quality scans are not even necessary to yield results; a simple webcam snapshot of an image is said to work just as well because the search is based upon the patterns contained within the image.

If such technology catches on, the rules surrounding fair usage of protected photography on another's Web site without express permission could tighten.

TinEye is taking requests for participation in the private beta through this link.