Using Facial Recognition
Improving operations and security in the airport setting
Airports are unique places. They are accessed by
thousands of people every day. Few other facilities
in the world have to screen so many individuals
for personal and environmental safety while
providing a positive experience. Further, the challenge
is not static. As new threats and requirements emerge, the
security challenges associated with airports continue to evolve.
Fortunately, advances in artificial intelligence mean airport
personnel are now able to leverage an asset they have in great
number to respond: security cameras. By bringing facial recognition
and other forms of computer vision to these cameras, airport
leaders are able to identify security threats much earlier than before,
as well as opportunities for improving operations that previously
required much greater effort and cost. In terms of security,
these same solutions provide a clear picture of who is coming
through their facility by instantly detecting and matching faces
to watchlists containing persons of concern and missing persons
in real time.
For operations, today’s AI-powered facial recognition solutions
for live video give insights into how individuals are moving
through the space and enable much faster access.
With access to greater understanding, airports are able to
make data-driven decisions that augment their security infrastructure,
enhance passenger experiences, increase operational
efficiency, and streamline after-incident investigations.
SECURING THE FRONT DOOR
Airport security advancements have focused on keeping passengers
and flight crew safe on airplanes. Strict policies and improved
scanning technologies have been implemented in airports around the world to mitigate the risk of in-flight terrorism. While these
programs have been successful, there remain other security challenges
that are not yet being fully addressed. In fact, when looking
at aviation terrorism more generally, research shows that there
has been no decrease over the years in the frequency of ground
attacks at airports.
Consider the 2016 attacks on airports in Belgium and Istanbul.
The Zaventem airport in Brussels was attacked on March
22, when suicide bombers detonated explosives in their suitcases
while standing in the check-in line. Three months later, on June
28, gunmen opened fire at an X-ray scanner in the Atatürk airport
before detonating the explosives they were wearing.
In both cases, the attackers took advantage of an existing
lan side vulnerability, namely the fact that airport security essentially
begins at the checkpoint well inside the building. Until
a passenger is processed through that first point, there is no way
for security staff to know who is coming through their front door.
The challenge of securing the perimeter is one that SAFR is
helping airport operators and law enforcement address by augmenting
existing video surveillance systems with facial recognition
technology. For airports, facial recognition technology can
help manage watchlists, alert personnel to unauthorized individuals
in secured areas, and locate and reconnect lost family members.
When it comes to responding quickly to terrorist attacks, facial
recognition can support security staff by providing them with increased
visibility and situational awareness. When you consider
that today’s airport security systems extend all the way to the
edge of the parking lot, you can understand the potential benefits
of facial recognition technology. If that system is able to identify
and match potential threats to watch lists and other databases, it
can prevent serious incidents from happening.
Facial recognition can also help prevent future attacks when it
is integrated into the post-event investigative workflow. By working
at incredible speeds to quickly identify anyone who attackers
have interacted with, this technology can provide investigators
with new insights as well as identify possible collaborators and
other persons of concern.
For example, during the investigation into the November 2015
attacks in Paris, the authorities were able to identify collaborators
using facial recognition and prevent further acts of terror.
WITH SPEED AND CONFIDENCE
If an incident occurs, it is crucial that airport security be able to
quickly identify and find the individuals involved. This is where
facial recognition can make a big difference. Working with AI allows
security personnel to search for identified individuals both
historically and in real time.
When combined with traditional video surveillance, facial recognition
technology allows security staff to ask the system for
all video clips that include a person of concern. Being able to do
this across multiple video feeds can provide a rapid view of everywhere
someone has been within the airport.
The ability to track and locate persons of concern quickly
enables security personnel to respond decisively. Depending on
the nature of the event, this can help stop or contain an evolving
Once an incident is over, security personnel, law enforcement,
and others shift their focus to investigation. At this point, facial
recognition technology can provide a clear picture of where persons
of concern have been, who they interacted with, and how long they were in the airport. This can help investigators understand
what led up to an event, who was involved, and if further
investigation or identification of collaborators is required.
Seamless Curb-to-Gate Access
Airport function has expanded. Once seen as purely functional,
airports now focus on revenue generation across range of economic
activity beyond flights, including tailored consumer experiences.
Every minute a traveller saves from a faster line, is another
minute they are able to enjoy an airport’s retail offerings.
When passengers don’t have to stand in line waiting to checkin
or get their boarding pass, they are less frustrated and spend
more time—and often more money—in restaurants or at the duty
free. Because they are one of the primary revenue streams for airports,
offering more and better services makes economic sense.
To ensure an easier journey, airports are now considering biometric
technologies. The idea is that, in essence, your face could
be your boarding pass. You could walk up to a kiosk, be recognized,
and then be directed to your gate—100 feet to the left.
While this technology has not yet been fully implemented,
there are currently a number of airports around the world that
are undergoing major renovations to be able to provide this experience.
And we are excited to be helping airports along this path.
PROCESSING PASSENGERS FASTER
In 2019, SAFR was one of 12 solutions accepted into in a Biometric
rally held by the US Department of Homeland Security. The
rally was a hands-on competition based on a set of pre-established
performance criteria about how quickly you could process a passenger
through a check-point. There were two components: (1)
getting an optimal image of the face for facial recognition and (2)
matching that face against the database. SAFR (code-name Jarvis)
came in second, well ahead of the industry’s largest players.
When it comes to facilitating curb-to-gate access, having capabilities
around rapid biometric acquisition and matching will
be crucial for providing the seamless passenger experience that
airports want to deliver. Faster and lower-frustration traveller
journey is good for travellers, airports and airlines.
INCREASE REVENUE THROUGH
IMPROVED OPERATIONAL EFFICIENCY
How long does it typically take passengers to move through airport
check-in and security? Does that vary depending on the time
of day or the day of the week? Unfortunately, most airports do
not know how long a multi-stage journey through their facilities
With this information, airports could begin to make operational
decisions that keep passengers moving and enhance their
overall experience. Reducing bottlenecks can have a positive impact
on how passengers feel. It can also improve security by reducing
the places where incidents can occur as well as providing
personnel with a less cluttered view.
Until now, airports have relied on people-counting applications to get a general sense of where people are at a given time. But,
they don’t provide the full picture. With facial recognition technology,
airports have access to a wide range of data points as well as
a more precise view of what is happening. This can be especially
important when you are looking to understand how quickly people
move from one place to another.
Before AI-powered facial recognition, you had to extrapolate
how quickly people moved from point A to point B by calculating
how many people a camera at each location captured within a
given time frame. But the system didn’t know how long it took an
individual to get from one place to another. Using facial recognition,
you can calculate how long it takes a specific person to move
through the space.
CAPTURING THE REAL PASSENGER JOURNEY
The challenge here is initially capturing a high-quality reference
image of an individual at the start of their journey so that the AI
can then identify them again in a second camera later on. Given
our world-leading accuracy and performance on live video, we
are enabling customers around the world to do just that.
To illustrate, when a face passes by a camera, our platform creates
a biometric template, or signature, that is stored temporarily
along with the time of day and camera information. When that
individual’s face is captured again by another camera later during
their journey, the image is then compared to the previously stored
face templates. A match between the images then provides the
exact duration for that traveller’s journey.
Of course, one of the concerns around facial recognition technology
is the perceived invasion of privacy. At SAFR, we understand
this and built Privacy By Design into SAFR from the start.
For this use-case, the identity of the traveller is kept anonymous
as the process does not require matching their face against any
databases. Individual privacy is further protected since all face
images and templates are automatically deleted from the system
at pre-determined intervals, usually between two and 24 hours.
IMPROVING AIRPORT OPERATIONS AND SECURITY
It’s clear that airports are complex environments that require
powerful solutions to address their unique needs. They have a
broad range of challenges relating to operational efficiency and
security. And, with so many assets, systems, and people to safeguard,
there’s little room for error.
Our goal at SAFR is to provide the technology that gives airports
the actionable data they need to enhance passenger experience
and maintain safe environments. With more data and a
greater understanding, they can make better informed
operational and management decisions
while also mitigating risks and responding
to events and incidents more effectively.
This article originally appeared in the May/June 2020 issue of Security Today.
Dan Grimm is the vice president and general manager of Computer Vision at SAFR.