Surveillance just added another weapon to its growing arsenal; identification by wobble, or as described in the recently published paper in Egocentric Video Biometrics “a person’s gait.”
Using data compiled from videos created by GoPro cameras mounted onto the helmets of 34 different subjects, researchers Yedid Hoshen & Shmuel Peleg of Cornell University were able to identify unique signatures in the differentiating wobble from just four seconds of camera footage. This, they say will compromise ego-centric (mounted) camera wearers anonymity, although it could have some benevolent uses. Your newly purchased camera can be tailored to recognise only your movements which may prevent some thefts, or user analytics on video sharing websites.
The experiment has only so far been performed with baseball cap mounted Go-Pro cameras but researchers plan on expanding the tests to include Google Glass and body mounted surveillance cameras such as those soon to be in use after the order of 50,000 units for the US police force was approved.
Perhaps we can finally learn the truth behind the Italy Go-Pro camera robbery in which an armed robber enters a supermarket and terrifies the public whilst looting. If you’ve not seen that, you’re in for a thrill:
We’re one step closer to Completing the Circle, although I have to admit to not considering this method of surveillance before. Scary stuff or much needed improvements in tech? Leave your comments below.
h/t The Verge
3 thoughts on “The ‘Wobble’ Made Whilst Holding a Video Camera is as Unique as a Fingerprint”