TinEye web site: http://tineye.com/login
This was taken from Ars Technica:
Another day, another new search engine makes the rounds. This time, it's an image-based search engine called TinEye, which has recently been opened up to the public in beta form. TinEye claims to do for images what Google does for text, which is to find web pages containing a specific image that you supply. This may not seem very useful for the average person upon first blush—after all, you have to already have a copy of the image in order to search for it—but could prove helpful to photographers and those who want to monitor the use of their work across the Internet.
According to the TinEye FAQ, the search engine uses "sophisticated pattern recognition algorithms" to analyze an image pixel-by-pixel, and creates a fingerprint from it. From there, it searches across the web for other images that are similar to your image's fingerprint, and can even identify images that have been cropped, watermarked, or altered (to some degree). In order to search, you must either upload an image or point the engine to a URL of one. You can also use a browser bookmarklet or extension, which will let you hit a button any time you want to use TinEye to search for other images like the ones on the page you're looking at.
We took TinEye for a spin to see exactly how it worked and whether it was accurate in searching for the images we chose. Our first complaint is that TinEye requires you to register an account before using the site. No other search engines require this, which makes that step particularly obnoxious—but then again, TinEye isn't like other search engines. Indeed, we tried to upload random personal photos of things that weren't likely to be duplicated anywhere else across the web, just to see what would happen: a chinchilla, cats, shoes, an old iPod, sea lions in San Francisco, homemade paella... not a single one turned up any results, and that's not surprising. You're not supposed to use TinEye to casually search for things like you would on Google.
When it comes to matching up images that could be posted on other parts of the web, however, TinEye excels. I tried uploading the full, uncropped image of myself used in my author profile here on Ars, and it was able to find the correct author page in a fraction of a second. (Unfortunately, that photo was not posted anywhere else—the world will never know about my awesome glasses!) However, there's one photo of mine that I have seen gratuitously ripped off across the web for years now: a scratched-up first-gen iPod nano that we had destroyed for the original iPod nano review. I uploaded that puppy into TinEye, which spit out seven results.
Through TinEye, we have already found at least one new site that is using one of our images for commercial purposes without permission! However, these seven websites aren't the only ones that have reposted our image—trust us, we know. TinEye defends itself by saying that the number of pages and images it has indexed is still relatively small at 702 million pages.
TinEye CEO Leila Boujnane points out that the site isn't just limited to hunting out copyright violators; it can help find the origin of an image too. "It's being used by researchers who need to find where an image came from to provide attribution, even people who are trying to find out who people are in old photos," Boujnane told PC Pro. "We had somebody who had a photograph of a soldier who'd arrived on the beach at Normandy and they couldn't find their name. They did a whole bunch of searches on TinEye and found a tiny little photo on an American website that listed everybody who'd gone to Normandy with a photograph. That's exactly when TinEye is useful, when you have an image but no words."
Overall, we liked TinEye. Our opinion of the site transformed from skeptical to impressed as we used it more and more. It's obviously not meant for casual web browsing or finding out what that crazy animal you saw on the side of the road is. However, if you are trying to keep track of how your creations are used across the web, or you saw an image and want to find out where it came from, TinEye is incredibly useful. As it continues to index more pages—the company claims it's already indexing hundreds of millions more per month—TinEye could eventually turn into a popular tool.