Ask a group of 100 people how they search for images online and the chances are that the vast majority will give the standard 'Google' response. Indeed, Google Images is a useful search engine if you are satisfied with a scattergun approach to finding all images that match a specific search term. But what if you wanted something a little more focussed, if you'll excuse the photographic play on words? What of you wanted to find all the images taken by your camera and that now appear online?
Impossible, you might imagine, but actually it's not only possible but actually incredibly easy using a free new resource that has entered a public Beta. Called, with a sad lack of creativity, the GadgetTrak Camera Serial Search it does pretty much what it says on the box. Enter the serial number of their camera and search for images online that were taken by that particular camera. As I write this, there are some 2,923,272 camera serial numbers indexed. When you connect to the site you can watch the serial numbers, along with the make and model of camera they belong to, whizz past on a scrolling display under the search box. Hover your mouse pointer over that scrolling display and it stops dead, allowing you to click through and return a search for all the images taken by that particular camera that are now online. Great fun for the voyeur in us, but not the primary function of this resource.
That, according to GadgetTrak founder and CEO Ken Westin, is to "assist in the recovery of stolen cameras" as well as working with law enforcement agencies regarding copyright issues. Westin has even suggested the technology could be used to assist in solving crimes such as the distribution of child pornography.
Talking of technology, the system is made possible courtesy of two things: the EXIF tags in the photographs themselves, and a custom built distributed computing infrastructure to index and map them all.
The Exchangeable Image Format (EXIF) tags within each photograph record the serial number of the camera it was taken with (although this is dependent upon your particular camera model and not all cameras do this) as well as data including shutter speed, exposure, resolution, date etc. This is true of digital cameras and smartphones, and it is increasingly common for location-based data to also be included these days. Obviously, to be able to index and map the vast number of photographs uploaded takes some serious computing power. GadgetTrak uses a custom built distributed computing model to spread the load currently hundreds of disparate computer systems. So effective has this been that in the first week of Beta deployment some 2.5 million images were added to the database and 50,000 camera serial numbers are being added every hour of every day.
Currently GadgetTrak only works with a handful of popular photo sharing sites and does not support smartphone images. However, it is a Beta and more resources will be added as the system develops. It's well worth checking out in the meantime, and it certainly does take the realm of search into a whole new different direction...
