The history of computer-aided facial
recognition dated back to the 1960s, research in automatic face recognition
started back then. And in recent years significant progress in this area have
been spotted and a number of face recognition and modelling systems have been
developed and deployed (Stan Z. Li, 2011).


The first faces classifying system was
developed by a man named Woodrow Wilson Bledsoe in the 1960s. He created a
device that allowed its users to input horizontal and vertical coordinates of
facial features of the portrait on a grid, using a stylus that emitted
electromagnetic pulses and these metrics were inserted into a database. When
the system was given a new photograph of a face, it was able to retrieve an
image most similar to that face form the database (Robert S. Boyer, 1995). Such impressive
progress was achieved in the 1960s, but unfortunately, due to the lack of
technology and computer processing power, the progress of facial recognition
was halted.


Accuracy and
performance have been added into the solution as processor power increases over
the years, doubling every eighteen months (Bajramovic, 2013). Other than basic
facial features (like eyes, nose, cheekbones) many different aspects like skin
texture, skin tone, reflexion of the light, facial expressions, etc. are taken
into account as technology escalates, which means that finer details can be
detected and analysed.


In modern facial
recognition, there are two main types of algorithms used in recognition, the
geometric approach, which explores distinguishing features and the photometric
approach, which maps an image into statistical values and compares the values
with samples to eliminate variances.



Related Work

Many uses of
facial recognition have been seen both commercially and domestically in the
recent years.


Google has developed an update
for its photos application, which detects the user’s pets in their photos and
groups them together (Google, 2018).

What Dog, a website that is
heavily supported by Microsoft, allows its users to upload their dogs’ photos
and it analyses the photo and returns the breed of the dog (Microsoft, 2016).

Airports are improving travellers’
convenience and security by using facial recognition enabled camera to detect bad
guys. Biometric gates are used by immigrations and airlines to avoid illegal
entries and to speed up passport control queues (Brandom, 2017).

Social media like Facebook uses
facial recognition to allow its users to tag names to their photos and link
them to other users (Facebook, 2018).

Some streets in China has
installed a public shaming system to discourage dangerous behaviours on the
road. A set of facial recognition embedded cameras are adopted to find those
pedestrians who break the law by jay-crossing the road, their faces are shown
on a big screen close by for every one to see (France-Presse, 2017).

Retails uses facial recognition
techniques to assist their business, the facial expression of customers are
detected and evaluated. So that better planning can be accommodated accordingly (West, 2017).



The unlocking system and the
security of the payment system Apple iPhone X are built on the foundation of
facial recognition. In order to locate the user’s facial features, invisible
light beams are emitted from the front camera onto the user’s face, a 3-D mesh
of the face is created and measurements of this mesh are calculated. If the
system agrees that this face matches with the one in the database, access is
granted (Apple, About Face ID advanced technology,

Bistro is developing a smart
pet feeder that separates different pets in the house to prevent over-feeding.
The feeder recognises the faces of pets in the house and only gives out a
certain amount of food to each pet (Nuwer, 2014).

Home-made versions of a smart
kitty door is gaining its popularity over the internet, to only allow access to
a pet that belongs to the house and/or to prevent a cat bringing a certain
creature into the house (Vaas, 2017).

A student has designed a
walking cane for the blinds, which can detect faces up to 33feet away, and
would notify the user if the cane recognises anyone the user knows by matching
faces with the ones stored in the database (Alvarez, 2015).



Literature Review

Deep concerns have been raised on the matter
of privacy. Even though going through a queue-less border control at the
airport may be viewed as improvement on convenience and efficiency, but for
those of us who are fully aware of our privacy, it is an intrusive form of surveillance.
“The biometric databases that the government is amassing are simply another
tool, and a very powerful tool of government control”. “We can change our
bank account numbers, we even can change our names, but we cannot change our
faces and once the information is out there, it could be misused” (Khalid, 2017). Voices can be heard
from the cautious ones questioning not the system but rather how the data
collected from the system is stored and used.


Current use cases and theoretical
possibilities are largely discussed due to the technology’s potential of
becoming a substantial privacy issue under the EU data protection law. For European
retail communities to get around this, there are measures that traders can
take to comply with data protection law, including by means of information,
consent, and anonymization (Lewinski, 2016).


Although many reports have claimed that
there is a huge concern of privacy and worries of leak of personal data related
to the facial recognition technology. The level of security it provides has
reached a new high. Businesses are happy that shoplifters and known criminals
are identified with this system (Frey, Revealed: how facial recognition has
invaded shops – and your privacy, 2016). Residents of a
smart apartment that uses facial recognition to allow access are happy to see
the convenience it brings to their lives and the degree of security it provides (Denyer, 2018).

Alerts were also raised as to whether this
gives the government / authorities more control over our lives. Surveillance
technologies are giving the government a sense that it can finally achieve the
level of control over people’s lives that it aspires to” (Zenz, 2018).


It is said that smart systems embedded with
facial recognition face ethical problems of error, function creep and privacy. Human
errors are inevitable but Philip Brey, a professor of philosophy of technology at
the University of Twente suggested that the solution is all about finding the
right balance between security and privacy (Brey, 2003).


Modern law needs to be updated alongside
with this newly developed technology, as it crosses the line of privacy and it
is currently deemed as unlawful in practice in some places. Regulations need to
be put in place so that we can benefit from facial recognition technologies
with the high level of efficiency and security but also be protected by law
with our privacy and personal data (Dept. of Comput. Sci. & Eng., Notre
Dame, Univ., USA, 2004).

Written by