The described pixel attack works in the digital domain (ie. modifying
pre-captured images), in other words the attacker must have access to
the digital pipeline. For real-time applications such access is rarely
available, certainly not for the regular people on the street.
However, there are analog countermeasures:
https://arxiv.org/pdf/1602.04504.pdf
https://cvdazzle.com/
https://io9.gizmodo.com/how-fashion-can-be-used-to-thwart-facial-recognition-te-1495648863
I wonder if this will became a hoodie of the 21st century.
"CV Dazzle explores how fashion can be used as camouflage from
face-detection technology, the first step in automated face recognition.
The name is derived from a type of World War I naval camouflage called
Dazzle, which used cubist-inspired designs to break apart the visual
continuity of a battleship and conceal its orientation and size.
Likewise, CV Dazzle uses avant-garde hairstyling and makeup designs to
break apart the continuity of a face. Since facial-recognition
algorithms rely on the identification and spatial relationship of key
facial features, like symmetry and tonal contours, one can block
detection by creating an “anti-face”. "
https://arxiv.org/abs/1710.08864
One pixel attack for fooling deep neural networks
Jiawei Su, Danilo Vasconcellos Vargas, Sakurai Kouichi (Submitted on 24
Oct 2017)
# distributed via <nettime>: no commercial use without permission
# <nettime> is a moderated mailing list for net criticism,
# collaborative text filtering and cultural politics of the nets
# more info: http://mx.kein.org/mailman/listinfo/nettime-l
# archive: http://www.nettime.org contact: nett...@kein.org
# @nettime_bot tweets mail w/ sender unless #ANON is in Subject: