Searching for Your Photos with Tineye

If you post images on the Web, either as an artist, photographer or just a blogger, you realize that it is trivial for others to take those images and post them elsewhere. Many license their images specifically for that purpose while others attempt to prevent it as much as possible.

However, tracking your images once they’ve left your site is a notoriously difficult task. Though Google searches and digital fingerprints can help track text works, there is no easy way to search for image matches. Image searches still rely on text around and within the image and can’t actually look at the information in the picture to find duplicates. The best most can manage is just guesses based on keywords.

However, an image matching is looking to fix that. A company called Idee has released a new search engine called Tineye that seeks to make it both easy and free to find matches to your image on the Web.

Though the service has its limits, the technology has a great deal of potential and any visual artist who is interested in seeing where and how their work is used should consider giving it a try.

How To Use Tineye

Tineye does not work like other search engines, rather than looking for words or phrases, it needs a reference image. You can provide it one by either uploading it or providing Tineye with a URL to locate it.

Once the image is uploaded, Tineye automatically begins searching for any potential copies of the work. After a few seconds, it displays the results

For example, running this version of the Google logo:

Produces this match:

However, what makes Tineye so powerful is that it is not only capable of detecting direct matches, but also copies of the work that have been altered. If an image has been re-sized, cropped, scaled or even had its color altered, it can locate matches and will display them lower down.

For example, one of the lower results for the Google logo above was this:

This can be useful in cases where an image was heavily modified to fit in with another site, such as cropped for an advertisement, had its color altered to fit with a theme or otherwise changed to serve a purpose.

Though the power of Tineye is pretty obvious, it does come with a few limitations that need to be discussed before anyone gets too excited about the service.

Drawbacks and Limitations

The biggest problem Tineye has at this moment is the size of its database. According to a recent update, they estimated the size at over 1 billion images. Though that seems like a very large volume of images, Photobucket alone has well over 5 billion images and Facebook has an estimated 10 billion.

Obviously Tineye is not tracking most of the photos on the Web and, looking through the results, it seems to focus more on blogs and not on photo sharing Web sites.

Still, given that the service is free to try, all one has to do is register for an account, it seems worthwhile to give it a shot and run a few images through. Even if it doesn’t find every use, it is still better than most of the free alternatives out there.

Other Alternatives

Though Tineye is more than adequate for the curious or those who don’t find it important to track every single use, professional artists, photographers and stock image agencies have a stronger need to be aware of every use as it is potentially a lost licensing fee.

In those cases a more powerful service is needed. Picscout, for example, provides a similar image matching service to Tineye but works by indexing your works and then matching them against what they find on the Web. This way, it does not need to keep a large database of images, though it will also not return all of the results at once.

Another company, Digimarc, has produced an invisible watermark that it has the capability of tracking. To use this service, you simply apply the water to the image before uploading and, though it is invisible to the user, it can be detected by their spider, even if the image is altered.

Both of these services are for a fee and will likely be out of reach for most small bloggers and artists. Still, for those with greater needs in this area, they are very natural choices.

Conclusions

Image copy detection, especially for amateur artists, is still a budding field. The technology to make it cost-effective and powerful enough for every day use is really just reaching fruition outside of the high-to-middle end markets.

These next few years are likely to be exciting times for this field with new services cropping up all around. That is likely a very good thing as new legislation, especially the orphan works bill, could make such technology more important than ever.

If there’s one area where a better detection system is needed, this is it.

Written by Jonathan Bailey from Plagiarism Today
Posted on January 15th, 2009 and filed under Blogging
Do not forget to subscribe to our RSS feed for updates

3