BERLIN (dpa) – It’s possible to identify the person who took a picture or a video just by analysing the wobble of the image.
Researchers at Jerusalem’s Hebrew University made the discovery, German website Heise Online reports.
Whether it is a camera clipped to one’s clothing, a pair of Internet spectacles, or just a handheld camera, each of them shakes during filming because of the way the wearer walks or moves.
That wobble is as individual as a fingerprint, say the researchers.
The researchers could identify 88 per cent of 34 image takers based on a 12-second first-person-perspective video.
Of course, the system only works if analysts have earlier video images for comparison’s sake.
The scientists say their goal was to make clear that just because a video doesn’t have obvious identifying features, that doesn’t mean there aren’t details that can’t be detected from it.