Cool Tools
Login  |  Register

TinEye

TinEye is a handy reverse image search engine. If finds where on the web an image comes from. You can use it to find where a photo of yours appears elsewhere, to find a higher res version of an image, or to locate the origins of a photo someone forwarded to you.

It does not use keywords, watermarks, or inbound links (as Google does) to locate images; rather it locates images via matching digital fingerprints of the image's pixel arrays. This means it can find images that have been renamed, or cropped slightly, or even screen grabbed.

Currently TinEye is not exhaustive. In my experience it won't find all the copies of an image. (They only claim to recognize a billion images so far, which is a small subset of all images on the web.) But it will find enough to be useful.

It can browse your hard disk for a target image, but even cooler is the Firefox plugin which enables you to select in image on a web page and in a click find where else on the web this image also appears.

-- KK

TinEye

tinyeye2.jpg

NOTE: TinEye's business model seems to be using this technology in mobile phones to identify products. You snap a picture of an item and it gives you info about where to buy it, sample it, etc. Their first product (which I have not tried) is an iPhone App that recognizes CD covers (the few still being bought!) to provide you with the iTunes album options on your phone. If you have you used this App or anything similar and can report positively or negatively, please let us know below in the comments or via the submit page.

 




Rail Riders

Toys from Trash

Korg nanoKEY





Comments

 
#1 | Mon, 01-12-09 09:50
Leila Boujnane

Hi Kevin: thanks for the TinEye mention. Just a tiny correction: our current image index is 1 billion. This means that when you upload an image to TinEye we are searching through this 1 billion image index to find your image. Admittedly 1 billion is still a tiny sliver of the web but it is still bigger than 1 million! We are currently growing our index so expect a few billions soon.

Leila

 
#2 | Mon, 01-12-09 12:35
Kevin Kelly

@ Leila: Billions! Right. Fixed it.

 
#3 | Tue, 01-13-09 11:46
Dave Cortright

I've used TinEye occasionally over the past 6-12 months. It's great for finding a cleaner, higher-res version of an image. Especially one that some lame blog felt the need to watermark their logo over.

I even used it to find the highest res credit card logos out there, and packaged them up for other e-commerce sites to use. Amazing that the credit card companies themselves don't do this, but there it is.

http://www.kpao.org/blog/2008/09/hi-res-credit-card-logos-visa-amex.html

 
#4 | Wed, 01-14-09 05:20
Corey L.

Quite a neat service. I've needed something like this a few times in the past, and I'll keep it in mind whenever I may need it again.

 
#5 | Wed, 01-14-09 07:29
christopher

I'm actually looking for an image search that takes a description or sketch of an image and returns possible matches. For example, if I described or sketched the Dallas Cowboys helmet logo, it should return at least one hit on the helmet, among others because it's just a star.

This probably has little utility but I see these bumper stickers with two vertical yellow lines on a blue background, and I have no idea wtf it means.

 
#6 | Wed, 01-14-09 05:46
Megs

Mmmm, I wonder if this could be used to track down stolen art.

 
#7 | Thu, 01-15-09 07:57
Daryl Spitzer

The SnapTell iPhone app appears to use similar technology. I've only tried it once, but it worked perfectly: it identified a book in about half a minute (over EDGE) after taking a photo , and gave me several useful links to choose from.

on iTunes (free): http://itunes.apple.com/WebObjects/MZStore.woa/wa/viewSoftware?id=291920403&mt=8

 
#8 | Sat, 02-07-09 08:11
Peter

I love SnapTell. That is truly a cool tool. Are they related to TinEye?

I tested TinEye on my own photo that I used in two websites. Didn't find anything. Also, I did find many copies of a creative commons image, though not the source at flickr.

 

Leave a comment



Thanks for your comment. The words in the CAPTCHA box come from old book texts that are being scanned and stored by the Internet Archive. By entering the words in the box, you prove you are not a bot and also you help proofread the books. If the sample you see is too hard to read, simply click the recycle button to get another two. Don't forget to put a space between the words.