Today I came across an interessting website. It's a search engine but stay with me, here. TinEye does what they call "Reverse Image Search". Pretty interessting concept with a downside: the still very small index.
Since reading is exhausting, let's just watch the pretty lady explain it to us.
In other words, you can enter an image URL or upload an image and with the help of a "digital fingerprint" of the image TinEye searches its index for dublicates, crops and altered version. Pretty amazing.
Possible uses
Okay, now how can this come in handy?
- First of all, you can find out where an image came from that you found on the web.
- Got an image of the web and need a higher resolution? Test your luck.
- Afraid someone might be stealing your photos? Find out who.
- But also: You have released something as public domain or under a CC license. See who uses your work where and for what purpose.
- Last but not least, find altered versions of your pictures. See what other people make of it.
But does it work?
That's what I was wondering. The video shows a nice little demonstration of Mona, but would it work with my images? The technology must be somewhat alright if Adobe has licensed it.
Well, I tried it with some of my older photos. None of them were indexed. Then I tried it with my public domain atom feed icons. No luck either. One of them didn't turn up any results. The other quite a few, but nothing I had hoped for.
As this icon uses a fairly common symbol the search turned up quite a few results, but no dublicats of my icon but rather a bunch of other images using the standard atom icon. No hard feelings. Apparently the Twitter clone indenti.ca has not been indexed, yet (they use this icon e.g.).
I was impressed to see, though, that the images found were very similar and although they all used slight variations of the symbol they all were decent matches.
So what now?
Would I say this works? Yes, but... Since the index is still very small the results are fair but not great. But if we are to believe the lady from the video TinEye's busy crawling spiders are at it. So let's give it some time, shall we?
Related links
Trackbacks |
Trackback adress: http://click.blogage.de/trackback/add/13742
No trackbacks, yet.
Comments |
In fact, Tineye is water under the bridge. Searching for one image at a time is just too much expenditure of time. The new automatic search engines are much better: upload once, and they keep on searching.
Look at www.imagerights.com, that's the new generation.
sofias
cool stuff :)
yet another part of visual cortex of the internet :D