The Truth about Biometric Exit
Technology to verify departures by foreigners has become affordable
- By James Albers
- Oct 01, 2014
How can the United States ensure that the millions of visitors who
travel to this country with a temporary visa uphold their agreement
to leave before it expires, while also enhancing convenience
for travelers? Biometrics technology holds the answer. In fact,
this technology dramatically lowered cost estimates for U.S. airport
exit control systems from $6 to $3 billion in 2008 to $1 billion today.
However, despite more than 17 years since a law called for a biometric exit
system to help secure our nation’s air and seaports and more than 12 years since
the 9/11 Commission recommended it, we still don’t have a system in place in the
United States. The problem is that some people continue to rely on outdated cost
estimates and technology when considering the solution.
In reality, advanced biometric entry and exit control systems are already being
used in airports and border crossings outside of the United States. These systems
are affordable, convenient, and can accommodate a high rate of traveler throughput.
When properly implemented, biometric exit controls can provide a higher
degree of identity assurance than biographic controls alone. Furthermore, this can
be done without disrupting operations at airports, seaports or other ports of entry,
and at a reasonable cost when considering the benefits.
What is Biometrics?
While the latest iPhone familiarized us with biometric technology, particularly fingerprints,
most people still don’t have a full understanding of it. Biometrics is who we are; each mode of biometrics, including face, fingerprint
and iris, contains a number of unique identifying
features. They are completely unique to individuals and
cannot easily be faked. Each distinct biometric modality
offers numerous unique benefits, that, when used
together, can dramatically increase identity assurance.
Facial recognition. Though not a new concept, as
police officers have been using it for as long as there
have been criminals such as when reviewing mug
shots, algorithms for face recognition have progressed
to the point where software is better at matching faces
than the average human could. Face recognition is
a potential solution for biometric entry and exit as
high-quality video cameras are already installed at
most major airports. If mounted in the gates used for
international departure, these cameras could capture
the faces of departing visitors.
Iris enrollment and verification. In 2008, this was
still a niche technology, but today, iris devices have
become reasonably priced, as several iris camera manufacturers
have driven down the costs per unit with
high-quality image results. The recent addition of iris
to the NIST Personal Identity Verification specifications
has added certainty to the technology and gives
it room to grow.
An iris template cannot easily be reverse engineered
to the original iris photo; thereby, it protects the privacy
of the traveler’s biometric information. Iris recognition
can be unobtrusive. High resolution cameras have
improved to the point that photographs taken from up
to three meters away from the subject will have the desired
quality, along with the good focus that is required
for iris recognition. Similarly, iris recognition systems
are contactless and hygienic. There is no need to touch
a surface that has been touched by thousands of travelers,
helping to prevent the transmission of disease via
contact. Considering fingerprints can sometimes be associated
with criminal behavior, iris may be less offensive
to a subset of travelers.
Iris recognition systems are stable. The iris is an
internal organ that is protected against damage and
wear by the cornea. It is mostly flat with a geometric
configuration that is mainly dictated by pupil dilation.
This makes the iris shape far more predictable and
stable than an individual’s face. Because of this, iris
recognition is highly reliable with the ability to reach
extremely low false accept rates, comparable to that
of any other biometric modality.
Finally, iris results are returned quickly. Iris templates
are small, and as a result, database searches
are fast. Current server hardware can match an iris
template to tens of millions of templates in less than
a second. This means that iris can provide a “lights
out match,” comparing one template to many records
faster than any other biometric.
Stand-off dual capture. There is even more to gain
by using stand-off dual capture units, which simultaneously
captures both face and iris images. Today’s
dual face and iris cameras are of sufficiently high resolution;
a photograph taken with the appropriate lens
and lighting, from a distance, still has the resolution
necessary to perform accurate matches.
Stand-off technologies allow for photographs to
be taken unobtrusively, in seconds and can be configured
to fit into natural chokepoints at airports, such
as check-in counters, security checkpoints or at the
departure gate. Within seconds, a stand-off simultaneous
face and dual iris biometric collection device
can take photographs of a subject’s face and eyes and
send the images to a back-end search engine. Depending
on the complexity of the search workflows, results
can return in seconds. At natural document handling
points, search results could be returned before
the document examination is completed. With faster
computer systems and advances in computer vision
algorithms, these systems will only get better.
Fingerprint recognition. There are exciting advances
in fingerprint capture devices that provide new
options for exit configurations. There are contactless
fingerprint capture solutions that can capture four finger
images with a single movement of the hand in less
than a second. Its fast acquisition capability allows
subjects to provide fingerprints while on the move,
making it suitable for high-traffic environments. This
method of fingerprint capture alleviates hygienic concerns
when using a commonly touched surface.
Biometric Exit Offers Greater Security
The use of biometrics is the single best way to quickly
and accurately prove an identity. Biographic information,
such as a person’s name and date of birth, as well
as the documents used to verify biographic information,
are all vulnerable to fraud. All of this information
and documentation can be falsified and stolen. In addition,
biographic data is fraught with errors because
it is reliant, in most cases, on human collection. The
Boston Marathon bombing serves as an example of
this: Russia warned the United States about Tamerlan
Tsarnaev, but an opportunity to question him was
missed because our records contained a spelling error.
Biographic data is presented inconsistently around
the world. Birth dates can be presented as day/month/
year or as month/day/year and names can be presented
as first/middle/last or the reverse. This is in
addition to limitations associated with transposed or
erroneous numbers or letters. Many names—especially
those associated with foreign languages—allow
multiple spellings or use of hyphens or other marks
that can reduce the reliability of biographic systems.
Biometric identification and verification, on the other
hand, is based on international standards and is generally
not subject to the same vulnerabilities as biographic
data.
Multi-Modal Biometrics Takes
Security to a New Level
Today, when a nonimmigrant arrives at a U.S. port of
entry and applies for admission to the United States
by air or sea, the only biometrics collected are fingerprints
and photographs; but, the photographs are often
not suitable for use by facial recognition technology.
Because of this, many assume that an exit system
should be based on a fingerprint matching system.
Collecting additional biometrics from visitors, fingerprints
for sure, but also quality photos that work
with facial recognition systems and are iris-capture
compliant with recently issued National Institute of
Standards and Technology (NIST) standards would
provide additional flexibility to employ a system that
works most effectively in different airport environments.
In some instances, this may include iris scans;
in other instances, facial recognition; and in others,
fingerprints.
Regardless, collecting multiple modalities of biometrics
at the time of entry would provide more options
for capturing biometrics at exits. Contactless
biometric capture using face, fingerprint, iris scans or
a combination thereof would allow operators to take
advantage of the relative benefits of each biometric
identifier and method of capture such as accuracy,
passenger throughput, convenience and cost. In this
scenario, fingerprints would continue to be collected
during enrollment, allowing for comparison against
relevant databases. And, with the addition of iris and
facial recognition scans at the time of entry, the added
security benefit of allowing matching against face and
iris databases would be provided.
With multiple modalities implemented, the primary
modality used on a given day can be changed,
mitigating the ability to plan for methods to spoof
any single biometric. In addition, a multi-modal system
ensures that each visitor would have at least one
biometric identifier in the system that can be used to
confirm identity. And, while each of the three biometric
technologies has a relative margin of error, collectively
they can ensure high probability matches on
data sets with tens of millions of records and more,
and can be used to reduce the impact of failed biometric
capture, for instance, due to dry fingerprints.
Statistics from the Indonesian multi-modal national
ID card project, for example, shows that there is only
a 0.008 percent chance of false positive identification
and a 0.18 percent chance of false negative identification
on a database of more than 100 million records.
Additionally, multi-modal biometric entry/exit
systems actually reduce processing times. Combination
facial recognition and document scanning systems
can process a passenger in as little as eight seconds.
Iris scanning systems can capture and process
an iris image in a few seconds, while contactless, finger-
on-the-fly technology can read four fingerprints in
as little as three seconds.
The Cost of Biometric Technology
Continues to Drop
In studying the opposition to biometric exit controls,
there is significant reliance on a 2008 report that estimates
the costs for implementing a biometric exit system
at airports and seaports between $3 billion and
$6 billion. These cost estimates are out of date and orders of magnitude too high, and do not take into account the dramatic price
declines in biometric technologies in the six years since—the same kind of declines
we’ve seen in other technologies.
This original study rested on the assumption of building a system to specifications
without considering the range of commercially-available, off-the-shelf
(COTS) biometric capture devices offered by vendors today that are affordable,
highly accurate and are designed specifically to fit within an airport footprint. To
illustrate this point, the first live scan fingerprint devices were put into use 20 years
ago, were big and bulky, and cost $15,000 or more per unit. Today, live scan devices
are small and cheap enough to put on an iPhone, and law enforcement agencies
buy high volume, 10-print fingerprint devices for less than $1,500.
Smartphones are a good example of how the cost and size of cameras has declined.
An iPhone 5 has an 8-Megapixel camera, and the newer models even have
autofocus. Nokia now has a 41-Megapixel camera. When this study was done, a
41-megapixel camera was impossible to buy at any price. These remarkable changes
have allowed cameras to become ubiquitous, and have facilitated law enforcement
activities using facial recognition.
Likewise, iris recognition was in its early days about six years ago, but it is now
recognized as the most efficient and effective biometric when comparing one template
with many records. A prime example of the scalability of biometric enrollment
and verification is the Unique Identification (UID) program in India. There
are 425 million Indian citizens now enrolled in that national ID program, with a
goal of enrolling 600 million by 2014.
Biometric Exit Systems around the World
Australia has an eGate based on facial recognition and ePassports called Smart-
Gate. SmartGate has since been expanded to New Zealand and has processed a
combined 15.1 million passengers since 2007. Currently, there are over 150 eGate
systems within 24 international airports across eight countries, processing over
one million passengers per month.
Other major deployments include fingerprint-based systems in France and Indonesia;
iris recognition systems in the United Kingdom and the United Arab
Emirates; and facial recognition systems in Germany and the Czech Republic.
Many of these systems are not limited to ePassport holders of the host country.
Of the systems mentioned, holders of second generation American ePassports are
able to use Australia and New Zealand’s SmartGate and the French Parafe, and
the systems in France, the United Kingdom, Germany and the Czech Republic are
open to all second generation ePassport holders of the European Union.
Closer to home, a collaborative partnership between government agencies,
known as the Apex Air Entry/Exit Re-Engineering (AEER) project, is evaluating
a variety of readily available biometric applications in a Maryland testing facility
designed to replicate airport environments. This testing is squarely focused
on helping the U.S. increase screening capacity to accommodate growing traveler
volumes and identifying cost-effective biometric exit solutions to meet the requirements
of the law.
Considering the significant advances in biometric technology, dramatically
falling costs and successful implementations worldwide, it is clear that a fully
functioning biometric exit system is quite practical in airport settings without
disrupting legitimate trade and travel. Nearly 12 years since the 9/11 Commission
recommended it and 17 years since Congress mandated
it, biometric technology to verify that foreigners have left the
country when scheduled is more than ready for deployment in
U.S. airports.
This article originally appeared in the October 2014 issue of Security Today.