Skip to main content

How the police can trace photographs back to your camera

Forensic experts have long been able to match a series of prints to the hand that left them, or a bullet to the gun that fired it. Now, the same thing is being done with the photos taken by digital cameras, and is ushering in a new era of digital crime fighting.

New technology is now allowing law enforcement officers to search through any collection of images to help track down the identity of photo-taking criminals, such as smartphone thieves and child pornographers.

Investigations in the past have shown that a digital photo can be paired with the exact same camera that took it, due to the patterns of Sensor Pattern Noise (SPN) imprinted on the photos by the camera's sensor.

Related: Could these tiny robots be the future of computer interfaces?

Since each pattern is idiosyncratic, this allows law enforcement to "fingerprint" any photos taken. And once the signature has been identified, the police can track the criminal across the Internet, through social media and anywhere else they've kept photos.

Researchers have grabbed photos from Facebook, Flickr, Tumblr, Google+, and personal blogs to see whether one individual image could be matched to a specific user's account.

In a paper entitled "On the usage of Sensor Pattern Noise for Picture-to-Identity linking through social network accounts", the team argues that "digital imaging devices have gained an important role in everyone's life, due to a continuously decreasing price, and of the growing interest on photo sharing through social networks"

Today, "everyone continuously leaves visual 'traces' of his/her presence and life on the Internet, that can constitute precious data for forensic investigators."

The researchers were able to match a photo with a specific person 56 per cent of the time in their experiment, which examined 10 different people's photos found on two separate websites each.

The team concludes that the technique has yielded a "promising result," which demonstrates that such it has "practical value for forensic practitioners".

While the certainty of the technique is only just better than chance, the technology is pretty new, and the numbers could get a bit more promising in the future. And, like fingerprints, the SPN signature would likely only be a part of the case being brought against a suspect.

Read more: Introducing AISight: The slightly scary CCTV network completely run by AI