Helen of Troy may have had a "face that launch'd a thousand ships", according to Christopher Marlowe, but these days her visage could launch a lot more besides.
She could open her bank account with it, authorise online payments,
pass through airport security, or raise alarm bells as a potential
troublemaker when entering a city (Troy perhaps?). This is because
facial recognition technology has evolved at breakneck speed, with
consequences that could be benign or altogether more sinister, depending
on your point of view. High-definition cameras combined with
clever software capable of measuring the scores of "nodal points" on our
faces - the distance between the eyes, the length and width of the
nose, for example - are now being combined with machine learning that
makes the most of ever-enlarging image databases. Applications of the tech are popping up all round the world.
In
China, for example, fried chicken franchise KFC recently unveiled its
first "smart restaurant" that uses facial recognition to predict what
meal customers are likely to want, based on their age, gender and the
time of day, while payments giant Alipay is experimenting with "smile to
pay" tech.
In the US, medical technology company NextGate has developed facial
recognition that can identify patients and link them to their medical
records. In Israel, meanwhile, "facial profiling" firm Faception
even claims its technology can tell if you're a terrorist, extrovert,
paedophile, genius or professional poker player by analysing 15 details
of your face that are invisible to the naked eye. It then uses the information to determine your personality traits, with the firm claiming it has an 80% accuracy rate. And
Russian app FindFace lets you match a photograph you've taken of
someone to their social media profile on the country's popular social
media platform Vkontakte. In theory, you could track down a complete
stranger you snapped on the bus or train.
"There are many potential applications of robust and reliable facial
recognition technology," says Prof Josef Kittler from the Centre for
Vision, Speech and Signal Processing at the University of Surrey in the
UK.
"Uses include security and surveillance of business, identity
verification for business transactions, personalised treatment of
regular customers, and analysing a customer's reaction to displays for
marketing purposes."
The university is currently heading a £6m
($7.5m) collaborative research programme funded by the Engineering and
Physical Sciences Research Council to develop next-generation facial
recognition technology.
If the face fits
One of the principal drivers of the tech is the security sector. Carl
Gohringer, founder and director at Allevate, a facial recognition firm
that works with law enforcement, intelligence and government agencies,
says: "The amount of media - such as videos and photos - available to us
as individuals, organisations and businesses, and to intelligence and
law enforcement agencies, is staggering.
"We're well beyond the
point where all of it is usable or viewable by us as human beings. So technology will be applied that results in new and interesting
mechanisms of accessing, analysing, ordering, structuring and processing
this visual minefield."
Machine-learning algorithms can sift through this vast store of data
and improve as they go. Tools such as FaceSearch, from Vigilant
Solutions, analyse more than 350 aspects of the human face, enabling
suspects to be matched to a cloud-based database of more than 15 million
"mugshots". But Roger Rodriguez, director of public safety
business development at Vigilant Solutions, says this technology has
plenty of business applications too.
"Want to cater to those
high-spending VIPs when they enter your store?" he says. "Facial
recognition technology can send instant alerts when that VIP enters. "Or it could be used on cruise ships to check passengers back in after they've disembarked the ship for a land excursion." Businesses looking for a competitive edge see the technology as a "game-changer", he says.
Privacy backlash
But what price privacy? A
report by Georgetown Law Center for Privacy and Technology estimates
that about half of US adults - more than 117 million people - have their
images logged in a facial recognition network of some kind - a trend
civil liberties group the Electronic Frontier Foundation describes as "a
real and immediate threat" to privacy.
The US National Security Agency has been harvesting such images for years.
And authorities seem keener than ever to use the tech, citing security and law enforcement as the main reasons. It was recently revealed that the city of New York plans to install facial recognition tech on its bridges and tunnels to scan and identify people driving in and out. In
the European Union, such technology has to comply with the EU's Data
Protection Directive and, from May next year, the General Data
Protection Regulation. But Ruth Boardman, data privacy specialist
at international law firm Bird & Bird, says individual rights still
vary from one EU state to another. And the automation of security vetting decisions based on facial recognition tech raises serious privacy issues.
"In
some countries this is not permitted at present," she says. "In others,
like the UK this is permitted, as long as appropriate safeguards are in
place - for example, ensuring that anyone who believes this is in error
can ask for the decision to be reviewed." Reassuring perhaps,
but little comfort if you find yourself barred from a venue because some
facial recognition software has mistaken you for a modern-day Al
Capone.
No comments:
Post a Comment