Just as ballistic experts can trace bullet casings back to the gun that fired the shell, university researchers have devised a way to trace specific digital photographs back to the exact digital camera that took the photo.

Every original digital picture is overlaid by a weak noiselike pattern of pixel-to-pixel nonuniformity, according to Jessica Fridrich, an associate professor of electrical and computer engineering at Binghamton University, State University of New York.

Although these patterns are invisible to the human eye, the unique reference pattern or “fingerprint” of any camera can be electronically extractedby analyzing a number of images taken by a single camera.

“This technique can provide a proof that a given digital image came from a specific digital still or video camera,” Fridrich said. Thus, whenever a digital image is associated with a crime – such as child pornography or movie piracy –investigators can now provide crucial evidence linking those images to specific cameras.

An example of the ‘noise’ researchers believe to be unique to each digital camera, allowing digital images to be linked to the precise camera that shot the photograph. (Binghamton University photo.)

Fridrich said that means that as long as examiners have either the camera that took the image or multiple images they know were taken by the same camera, an algorithm she developed can extract and define the camera’s unique pattern of pixel-to-pixel nonuniformity.

“The defense in these kind of cases has often been that the images were not taken by this person’s camera or that the images are not of real children,” Fridrich said. Sometimes child pornographers will even cut and paste an image of an adult’s head on the image of a child to try to avoid prosecution, she said.

“But if it can be shown that the original images were taken by the person’s cell phone or camera, it becomes a much stronger case than if you just have a bunch of digital images that we all know are notoriously easy to manipulate.”

Like actual fingerprints, the digital “noise” in original digital images is stochastic in nature – that is, it contains random variables – which are inevitably created during the manufacturing process of the camera and its sensors.

Binghamton electrical and computer engineering researcher Jessica Fridrich has devised a technique able to trace specific digital photographs back to the exact digital camera that took the photo, much the same as ballistics can trace bullets to guns that fired them. (Binghamton University photo.)

“This characteristic virtually ensures that the noise imposed on the digital images from any particular camera will be consistent from one image to the next, even while it is distinctly different,” Fridrich said.

In preliminary tests, Fridrich’s lab analyzed nearly 3000 pictures taken by nine distinct digital cameras. She was able to link every individual image to the exact camera that took the picture with 100 percent accuracy.

Fridrich said law enforcement is aware of her work, but before it can be used in the field, more testing is indicated.

“Right now we are focusing on analyzing the reliability and mathematically describing the algorithm,” she said.

The main limitation of the technique is that it currently requires either the camera or multiple images taken by the same camera, and isn’t informative if only a single image is available for analysis.

The technology, however, is still in the development stage and is constantly being improved.

“We would like to make the method as reliable as possible and make it work for all types of sensors, including low-end cameras, such as cell-phone cameras,” Fridrich said. “Already, we have expanded it to video. Scanners remain to be done.”

Fridrich and two members of her team – Jan Lukas and Miroslav Goljan – are co-inventors of the new technique, which is also useful in detecting forged images.

Fridrich said the absence of expected digital fingerprint in any portion of an image provides the most conclusive evidence of image tampering.

Douglas Page writes about forensic science and medicine from Pine Mountain, California. He can be reached at