You might want to think twice before illegally filming the next concert you go to — at least using a body-worn GoPro or Google Glass. Even if we don’t realise it, we all move in our own unique way. It’s like a fingerprint based on motion, and now researchers at Cornell University have developed an algorithm that can analyse footage and determine who’s wearing the camera.
Using just 12 seconds of footage, the algorithm analyses the vertical and lateral motion in a video to identify a unique signature. And in lab tests involving 34 test subjects wearing a GoPro camera mounted to a baseball cap, the software was able to correctly identify who was responsible for each video about 88 per cent of the time. So there’s still some wiggle room for a lawyer to claim it wasn’t you recording that concert.
The research is particularly well timed given the debate over police wearing body cameras while on duty. The new algorithm can help improve accountability if an officer knows there’s a foolproof way to tie them back to the footage. It can also help law enforcement identify who’s behind a clip posted online showing illegal activities, or at least help them build a stronger case if they have a suspect in custody. [Cornell University via New Scientist]