Su Reviews Idee’s TinEye
Apparently, I (along with anyone who is anyone) should already have known about a gem of a beta product called TinEye by Idée. Apparently, it’s been out for a few months. Apparently, it is freaking awesome.
Apparently, people assume too much and don’t promote enough.
You know what they say: If you assume, you make an a out of Su and me. And believe me, Su doesn’t like being an A. So don’t do it.
If you think a product is great, promote it! Tell people. Write up an article. Whatever. But don’t say, “Gee, yeah, I really loved that last week when I used it for 5 minutes and now you’ve missed the bus.” Because that is what you are saying. To me. And to Su.
Su doesn’t even have a car. Now he has to walk.
So what the heck is TinEye? Is it that decorating show, you know, the one with the people? Hardly.
If you assume that the Tin Man has eidetic memory and a talent for finding pictures that look like each other, you’d be getting close to what TinEye does.
In a nutshell, TinEye leverages some proprietary tech that Idée has been working on for creating digital fingerprints of images. This technology is mated with a high-speed matching engine and a database of hundreds of millions of public domain pictures to give you fast image matching based on a source picture.
After using the product for a while, I have to say I’m impressed. Currently, with its sub-set of all the pictures out there on the internet, it can’t find a lot of pictures uploaded to, say, Flickr. But if you do a Google image search and then drop the URL of the image you are interested in, it will find that picture wherever it resides on the internet.
For example, I searched for a picture of George W. Bush and dropped the URL in TinEye. Here is the result (click image for larger size):
It found the image in 272 different locations on the internet! It also gives the size of the image, and lists separate links to the same source image by domain. I do wish they would give you the option to re-sort the results by different criteria, such as image size, file size or bit depth.
So, if you want to know if a particular picture appears somewhere else on the image (say for copyright reasons), this is a great tool to have.
I’m really happening this comes out of closed beta soon, however, if you are clever you may already have noticed that the FriendFeed link at the beginning of the article has a special URL you can use.