The research, conducted at Binghamton University in New York, analyzes the slight variations created by the image sensor in each camera to uniquely identify pictures.
The technology is being presented as potentially useful in nailing child pornographers. "The defense in these kind of cases would often be that the images were not taken by this person's camera," Jessica Fridrich, the Binghamton University engineering professor who oversaw the research, said in a statement. "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."
In technical terms, Fridrich and her co-inventors Jan Lukas and Miroslav Goljan found that every digital picture is overlaid by a weak noise-like pattern of pixel-to-pixel non-uniformity. As with fingerprints, that digital noise pattern will be consistent among all images taken from the same camera. In preliminary tests, Fridrich's lab analyzed 2,700 pictures taken by nine digital cameras, with what the Binghamton University statement said was 100 percent accuracy. On the downside, the technique won't work on analysis of a single photo. Investigators have to have either the actual camera or multiple photos to get a line on the noise pattern.
Pegging a photo to an individual camera essentially extends forsenic document identification techniques to the digital imaging area. Such methods have long been used, for example, to identify handwritten ransom notes, and to tie printing materials to typewriters by looking at the unique characteristics of slightly bent keys or carbon-film ribbons.