Screenshot showing Google’s display of EXIF metadata
Google’s display of EXIF metadata. Photo © Jeremy Nicholl 2011. All Rights Reserved.

Professional photographers have rarely looked kindly on Google. The search engine giant has long been the first port of call for image infringers, and the situation was only made worse by Google’s habit of stripping metadata from images, turning them into orphans.

But over the summer things have changed. In June Google launched Search By Image, and the following month began displaying EXIF data in their image search returns. The latter absolves Google of the orphaning charge, but it’s Search By Image that’s the real game-changer. Google have promoted SBI as a tool for image users to find images they might want to publish, and it could of course be used by such people to track down photo owners so that the image could be properly licensed. But its real value to photographers lies in its ability to track down images that have already been stolen and published on the web.

Previously image searching on Google was a very hit and miss affair. Enter some keywords, and depending what had survived Google’s metadata mangling and any changes by the publisher you might or might not find the published mage: more usually than often not. But with SBI you enter a low-resolution version of the image on Google’s servers and they search the web for matching images: if any are found they appear in the search returns. And since a visual search is more precise and doesn’t rely on matching keywords SBI invariably returns far more results than a keyword search.

Of course SBI isn’t the first reverse image search: TinEye, PicScout and others have been running for several years. But Google’s offering is immensely more powerful simply because they’ve already indexed far more of the web than their rivals. Not that SBI is comprehensive: TinEye for one sometimes finds images that Google doesn’t, so ideally one should use several services when searching for infringements. But all of these searches are time-consuming, and SBI finds so many more images than the competition that if you’re only going to use one reverse image search then SBI has to be it.

All this is very bad news for photo thieves. For years they’ve heisted images with impunity: the chances of being caught were low, and if caught the default defence was “I found it on Google and didn’t know who the owner was”. But that excuse disappears with the now easily viewable metadata, and the chance of being caught has moved from remote to likely, going on inevitable.

Faced with the near-certainty of being caught a smart thief would stop. Unfortunately most image thieves aren’t smart, but dumb and lazy: too lazy to create anything original themselves, and dumb enough to think they can get away with stealing indefinitely. They’re also stuck in the habit of thieving, encouraged by spurious and misleading advice on such nebulous concepts as “fair use”.

So instead of the logical outcome of Google’s changes – contact the photographer for permission or don’t use the image – the actual result is likely to be lots more legal action from photographers. And of course lots more squealing from thieves who get caught.

But that’s their problem. The first reaction of photographers who try SBI is generally fury at learning the previously unknown extent of the theft of their work. That’s hardly surprising, but really we should be pleased to see the tables being turned, albeit slowly.

Bookmark and Share

6 Responses to “How Google Search By Image Helps Photographers Catch Copyright Crooks”

  1. HeruLS HeruLS says:

    Yeah, many photogs need this tool

  2. Pete Jenkins Pete Jenkins says:

    Hi Jeremy,

    My own extensive experiments with this and the other tools you mentioned suggest that this is still a hit and miss affair.

    If the image has had its exif data stripped, as is commonly the case with ‘orphans’ then there is no data to show.

    Worryingly I have had many ‘similars’ turned up on searches which are not really close to the original.

    Worse still on some searches one gets totally unrelated images returned.

    Granted it is better than it was, but I believe we have a long way to go before we can call Google ‘savior’. 🙁

  3. […] to track down images that have already been stolen and published on the web."Here’s a link to the blog post on Google’s Search By Image feature. […]

  4. […] a link to the blog post on Google’s Search By Image […]

  5. Jeremy admin says:

    Hi Pete,

    It’s early & I’d expect Search By Image to improve, & I suspect the results may differ depending on where you’re based. However I’m very surprised to hear that your results are so different from mine. Are you sure you’re using Search By Image and not the older text-based Google Image Search?

    Whether Metadata has been stripped affects Image Search, but not SBI. The latter isn’t searching text but looking for visually identical images: metadata is irrelevant to SBI.

    If you are using SBI you should ignore the Visually Similar Images link, since that does return a lot of irrelevant images. The important things to follow are “find all sizes” & “Pages that include matching images”: those seem to find all the indexed copies of a particular image.

    Doing the above I’ve found SBI to be very accurate. The only similars it found that weren’t mine were if a photographer was standing next to me & shot a pic that had the same very distinctive feature in it. For example SBI returned another photographer’s photo of a Moscow demo because both his and mine featured the same very large banner in the foreground.

  6. Melly Potter Melly Potter says:

    Google image search is looking for images that are similar in colour and shape; the subject on the pictures can be radically different! this explains the results you obtained, Pete.

    I did search for my most popular pictures, just in case. Honestly, I find fascinating to see totally unrelated pictures that look similar to a computer! I haven’t found any stolen pictures but had a little scare: someone posted a VERY similar pictures to one of mine. I was sure it was mine! But I look more closely, I noticed few discrepancies. So me and that perfect stranger have taken virtually the same photo in the streets of London! Fortunately, the people in the picture are not the same! 😉

  7. dejan dejan says:

    You can try Exif photo search it search flicker images by exif data.

  8. I realize this article is a bit dated, but im sure everyone on this thread including the author needs to know about this search engine as well :

Leave a reply: usual Marquess of Queensberry rules apply.


(required: will not be published)