Like people, cameras have unique identifying features much like fingerprints. As such, a research team from Charles Sturt University is developing a method for matching images and videos to the device used to take said images and videos.

According to the research team, the hardware and software processing components of digital cameras are responsible for the unique traces left behind in every image and video taken on a device.

“Like the barrel of a gun will leave unique markings on a bullet that passes through it, which can then be used to match the bullet back to the firearm that fired it, we can find something similar in digital images," said Dr. Xufeng Lin, research lead.

"For example, manufacturing imperfections in the camera's sensor will leave behind a pattern of noise in an image that is unique to that particular device," Dr. Xufeng added.

To demonstrate, Xufeng and the team tested the approach on a database containing 5,400 images taken on over 35 different cameras. Instances where the camera’s “fingerprint” could be credibly measured, the team was able to match the images to the correct device at least 95 percent of the time.

Additionally, the team created an algorithm capable of sorting through other larger image databases and assembling together images taken by the same device.

Currently, the technology has been used to help Sussex Police in the U.K. and Interpol to match child pornography images to a suspect’s mobile phone as well as to help identify victims by analyzing several million images held in the International Child Sexual Exploitation image database.

"As camera fingerprint techniques like ours improve and get more sophisticated, I believe that they will play an even greater role in fighting cybercrime," said Xufeng.