When I first heard of TinEye, my first thought was "What exactly is a REVERSE image search?" Well it turns out it's quite clever and useful. If you've got an image, TinEye will examine the actual image contents (not just keywords, tags, and such) to find the source (or any other instances) of that image. Of course it doesn't have ALL the images on the web, it's currently at about 2 billion, but that's enough to make it already useful. There's a web front-end, plugins for most of the major browsers, and even an API for external tasks. How useful can reverse image search be? Let's find out.
What it's For
The website describes itself as follows:
TinEye is a reverse image search engine. You can submit an image to TinEye to find out where it came from, how it is being used, if modified versions of the image exist, or to find higher resolution versions.
That sums it up pretty well, but there are a few side effects of this process. For instance, if you like pictures of a certain type of flower, or variations on a particular internet meme, TinEye's ability to match similarities can bring you results that would otherwise be difficult to match by filename or keyword. If you're a digital artist or photographer, it also has the potential to let you see where your work has gone and what changes have been made to it.
TinEye can find not just your image, but derivatives that all share the same basic image. This can help you find the same pic even if it's been cropped, rescaled, or sometimes even pretty heavily 'shopped'. Chances are, the first thing you'll want to do is try posting an image to see what comes up. For testing, I'd recommend something that's popular and likely to have been modified several times, like the Mona Lisa or a well known meme.
If you click Choose File, you'll be able to upload an image file for TinEye to scan.
Once the search is complete, you'll get the results ranked in terms of closeness of the match, according to TinEye's criteria. Further down the list, or in the following pages, you'll find the results that are more likely to be modifications of that work instead of a close match.
It's able to recognize Bubble Girl even though her background has completely changed with each image. I could make my own variation on the theme, post it on this page, and once TinEye gets around to crawling MTE, you'd be able to match my image to the original.
If you end up using this service a lot, the web front end might not be the most convenient way to go. To streamline the process a bit, you could install the TinEye plugin for your browser (Safari, IE, Firefox, and Chrome).
This will bring up the results just like a web search. You'll notice in this example that TinEye found one other screenshot of the same application, and then returned images with a similar dotted pattern but no longer the same application.
The more I use TinEye, the more usefulness I see in it. The total database is still somewhat small by current web standards, but all signs indicate that it will only grow as the service matures. While image matching is not something I'm likely to need every day, there are certainly times where keyword searches aren't enough. When that happens I'll be glad to have TinEye.
Friday, April 8, 2011
Published by Jeroen Adolfse
An Article from maketecheasier.com