Anti-surveillance clothing obscures wearer from facial recognition tech
Wed 4 Jan 2017
A new concept technology called HyperFace which attempts to defeat facial recognition software has been created by Berlin-based artist Adam Harvey.
Inspired by animal camouflage techniques, the system involves printing deceitful patterns onto clothing and textiles. The patterns are designed specifically to play on how computers detect faces, confusing the software by throwing out thousands of false facial features.
Speaking to The Guardian, Harvey noted that his HyperFace concept seeks to overload ‘an algorithm with what it wants, oversaturating an area with faces to divert the gaze of the computer vision algorithm.’
To create the prototype, Harvey collated 47 different data points which academic and commercial researchers claim to be able to discover from analysis of a 100×100 pixel facial image.
Harvey explained that HyperFace could be printed on clothing or textiles to cover an individual or the surrounding area to achieve a disguising effect.
Alongside partners Hypen Lab, the artist is expected to provide further details on how HyperFace will progress from concept design into reality during the 2017 Sundance Film Festival later this month.
HyperFace comes as an extension to Harvey’s ongoing project, CV Dazzle, which aims to mess with facial recognition technologies using makeup and hairstyling techniques to break up facial features targeted in vision algorithms.
‘I think this project could change fashion designers’ and architects’ approach to modulating the way bodies appear or disappear into the background of a computer vision readable world,’ said Harvey.
Tech giants are increasingly interested in using facial recognition solutions to support various digital initiatives. Amazon’s recently announced physical Go stores are expected to be equipped with facial surveillance, while Facebook uses vision algorithms to automatically tag users’ photos.
With facial recognition technology set to infiltrate more aspects of our everyday lives, anti-surveillance tools could provide interesting privacy protection mechanisms.