I enjoyed playing around with the service, and I especially liked the examples using iconic images like American Gothic and Michelangelo’s The Creation of Adam. It’s fascinating to see all the different ways these images have been changed and used for various purposes, and it’s impressive that TinEye’s algorithms are able to identify these altered images as copies of the original work.

However, as interesting as I thought Tineye was, I didn’t actually find it much practical use in finding unauthorized copies of my images on other webpages. I searched for some of my photographs that were used on various pages that I knew about, just to test it out, and had very little success. Tineye acknowledged that its database was still relatively small but growing all the time, but I stopped bothering using it to look for copyright issues. I had better success looking in my Flickr referrer logs and just doing word searches on Google Images.

But recently I decided to give it another try, and was quite impressed with my results. I searched a few of my photographs and found several unauthorized copies on various websites. Nearly all of my photographs are posted on Flickr with a Creative Commons Attribution / Non-Commercial / Sharealike license, and it’s discouraging to find people downloading them and using them without the required attribution, and on commercial sites. I send a message to each site asking that they add the credit line or remove the image, and most people apologize and comply. A few don’t respond or don’t seem to understand the problem, and one person was quite rude. It seems that there are a lot of people who ought to know better, including bloggers, journalists, teachers and sometimes even my fellow librarians, who would never commit plagiarism and give sources for quotations but treat other people’s photographs as if they had no value other than as eye candy, unworthy of a simple credit line. It’s especially annoying to find my work used without credit on blogs whose content is marked All Rights Reserved.

(You’ll have to forgive the soapbox — I found a lot of my uncredited copies of my photographs floating around the web today, thanks to Tineye!)

Links:

• TinEye
• A Simple TinEye Tutorial

Here’s an example of one of my Flickr photographs. TinEye has correctly identified a copy of this posted with permission by Schmap:

How does it work? According to the TinEye FAQ:

Using sophisticated pattern recognition algorithms, TinEye creates a unique and compact digital signature or ‘fingerprint’ for each image that is added to the index. When you submit a search image to TinEye, its fingerprint is analyzed on-the-fly and compared to the fingerprint of every single image in the TinEye search index.

In other words, it’s magic.

This could be a very useful tool for finding unauthorized copies of photographs and other images, and I’ve seen people on the Flickr forums who have had success with this. Unfortunately, the database is just too small at the moment to be really successful with this, at least based on my testing. I searched about twenty of my most Flickr photographs, including several that I know have been used on other sites, and the Big Blue Bear photograph is the only match I’ve found. But this definitely has possibilities for copyright issues and other situations where you want to track how an image has been used. Update: Tineye’s database has dramatically increased in size, and I’m now finding a lot more copies of my pictures on other sites.

They have also used it for some cool searches showing altered versions of famous paintings and other iconic works — Hello Kitty, the Linux penguin, the Abbey Road album cover, etc. Here’s a widget showing different versions of the Mona Lisa: