Computational Intelligence - February 2017 - 27

genuine but issued using false birth cerA single-phase Doddington's detector assumes a
tificates. That is, fraudulent breeder documents can serve as a basis to obtain
cooperative traveler who provides his/her biometric
genuine ID documents. A fast and reliable
traits according to the special rules of machine-human
watchlist check is the core of traveler risk
interactions. In the two-phase approach, we start
assessment. This leads to the need for
embedding this technology into social
to use a Doddington's detector before the traveler
infrastructure, in particular, via bridging
reaches the acquisition station.
the gap between the watchlist technology
and forensics.
However, in the practice of e-borders, any extension of biofacial biometric is the privilege of most documents and govmetric modalities in traveler authentication and risk assessment
ernment databases for the creation of watchlists using physical
requires the creation of costly supporting infrastructure. The
and digital (virtual) sources. Motivated by this fact, our study
leads to the following key conclusions:
4) We demonstrated using the Doddington's metric that
there is always a risk of impostors among persons of
interest. In terms of security, this means that a machine
0.010
1.0
may mistakenly provide border crossing passage to a
wanted person.
0.005
0.5
5) It is well documented that mitigating and suppressing the
Doddington's phenomenon is possible via multi-biometrics. Unfortunately in watchlist applications, including, for
0
200
400
0
2
4
example, fingerprints in addition to facial traits, is often
(a)
(b)
an unacceptable solution, in particular, because no such
Sheep
Traveler-04882d61
Goats
traits (fingerprints) are available in the digital/virtual
Traveler-04882d62
Wolves/Lamps
world (social media and surveillance networks).
6) Our study suggests a benchmark framework for the
figure 9 the risks of traveler 04882 from the frgc database in the
watchlist check. For example, Table 4 is the challenge for
kl metric over doddington's landscape. (a): genuine score, and (b):
improving Doddington's detectors, and data in Figure 10
imposter score.
should be considered as an indicator for the development
of more efficient recognition and inference techniques.
Acknowledgment

'Sheep'

'Goat'

'Wolf/Lamp'

Phase I
Phase II
Accumulated

0.9271
0.9612
0.9315

0.8792
0.8743
0.6076

0.9279
0.8159
0.9432

'Goat'
Phase I
Phase II
Accumulated

0.0696
0.0333
0.0678

0.1178
0.1199
0.3919

0.0652
0.0000
0.0000

0.0033
0.0056
0.0007
Session 1

0.0030
0.0058
0.0005
Session 2

0.0069
0.1842
0.0568
Session 3

This project was partially supported by the Natural Sciences and
Engineering Research Council of Canada (NSERC) through
the Discovery grant "Biometric intelligent interfaces"; the Government of the Province of Alberta (ASRIF grant and Queen
Elizabeth II Scholarship); and Defense Research and Development Canada (DRDC), Canada Safety and Security Program.

'Sheep'

'Wolf/Lamp'
Phase I
Phase II
Accumulated

figure 10 comparisons of the efficiency of watchlist inference in
three operating sessions for the frgc v2.0 database.

References

[1] A. Jain and A. Ross, "Bridging the gap: From biometrics to forensics," Philos. Trans. R.
Soc. B, vol. 370, no. 1674, pp. 1-10, 2015.
[2] R. G. Cowell, "Finex: A probabilistic expert system for forensic identification," Forensic Sci. Int., vol. 134, no. 2, pp. 196-206, 2003.
[3] F. Taroni, A. Biedermann, S. Bozza, P. Garbolino, and C. Aitken. Bayesian Networks
and Probabilistic Inference in Forensic Science. New York: Wiley, 2014.
[4] D. Bigo, S. Carrera, B. Hayes, N. Hernanz, and J. Jeandesboz. (2012). Justice and
home affairs databases and a smart borders system at EU external borders: An evaluation of current and forthcoming proposals, Centre Eur. Policy Studies, no. 52, Dec. 2012
[Online]. Available: https://www.ceps.eu/publications/all
[5] Government Accountability Office. (2007). Data mining: Early attention to privacy
in developing a key DHS program could reduce risks, Tech. Rep. GAO-07-293 [Online].
Available: http://www.gao.gov/new.items/d07293.pdf
[6] J. Kephart, "Biometric exit tracking: A feasible and cost-effective solution for foreign
visitors traveling by air and sea," CMS, 2013.
[7] Frontex. (2012). Best practice technical guidelines for automated border control (ABC)
systems, Research and Development Unit, Warsaw [Online]. Available: http://frontex
.europa.eu/assets/Publications/Research/Best_Practice_Technical_Guidelines_for_
Automated_Border_Control_Systems.pdf

February 2017 | Ieee ComputatIonal IntellIgenCe magazIne

https://www.ceps.eu/publications/all http://www.gao.gov/new.items/d07293.pdf http://frontex http://www.europa.eu/assets/Publications/Research/Best_Practice_Technical_Guidelines_for_

Table of Contents for the Digital Edition of Computational Intelligence - February 2017