2. Scope
• Overview
• Technology, Standards, Limitations
• Facial Recognition
• Applications and Advantages
• Is Visionics & Biometrics the Future?
3. Overview
• Visionics is subset of Biometrics
• Biometrics is defined as the use of
anatomical, physiological or
behavioural characteristics to
recognise or verify the claimed
identity of an individual.
4. Overview
• Originally developed for high
security applications.
• Confirms the presence of the
individual rather than a token
5. Technology
• Retinal Scan and Iris Scan
• Fingerprint and Hand Geometry
• Signature and Voice Dynamics
• Facial Recognition
• Vascular Patterns
6. Technology
• Other biometrics systems proposed but never
brought to market - yet. Include the use of
the earlobe, a person’s smell, and a person’s
gait to identify an individual.
• Much work in biometrics research, driven
primarily by the military. DARPA for example,
sponsors much of the research in the US,
with “Human ID at a distance” being one of
their ongoing projects .
7. Technology - Industry
• ~ 150 Biometric Companies
• Veridicom dismantled in August
• 2001: 60%+ growth (post
9/11/01)
10. Technology - Industry
• Harris Interactive (consulting) +9/11
– “82% of Americans are willing to have their
fingerprints scanned for increased airport
security”
– “86% favor facial-recognition technology to
scan for suspected terrorists”
– (CNBC 9/19/01: Alan Dershowitz: “facial
recognition is better than racial profiling”)
11. Technology - Industry
Wall Street Journal 11/13/01
“The Airport of The Future”
“A special scanner scrapped in
Charlotte, N.C., identified people
by their iris but couldn’t detect
guns.”
12. Technology - Standards
• The BioAPI standard is intended to
provide a mechanism where by a
single application can utilise different
biometric approaches. It provides a
standard interface between the
application layer and any biometrics
system that provides BioAPI
compatible drivers or subsystems.
13. Technology - Standards
• Two other standards of note in the
area are the ANSI X9.84 standard,
which is about common templates for
common biometrics devices and the
CBEFF (Common Biometric Exchange
File Format) is trying to establish a
universal file format recognisable to
various applications.
14. Technology - Standards
CBEFF Defined Biometric Type CBEFF Type ID Static/Dynamic Identification
Suitability
Multiple Biometrics Used 0x01 - -
Facial Features 0x02 Static No
Voice 0x04 Dynamic No
Fingerprint 0x08 Static Yes
Iris 0x10 Static Yes
Retina 0x20 Static Yes
Hand Geometry 0x40 Static No
Signature Dynamics 0x80 Dynamic No
Keystroke Dynamics 0x100 Dynamic No
Lip Movement 0x200 Dynamic No
Thermal Face Image 0x400 Static No
Thermal Hand Image 0x800 Static No
Gait 0x1000 Dynamic No
Body Odor 0x2000 Static No
DNA 0x4000 Static Yes
Ear Shape 0x8000 Static No
Finger Geometry 0x010000 Static No
Palm Geometry 0x020000 Static No
Vein patter 0x040000 Static No
15. Technology - Limitations
Biometric Crossover Accuracy
Iris scan 1:10.000,000+
Retinal Scan 1:131,000
Fingerprints 1:500
Hand Geometry 1:500
Signature 1:50
Voice 1:50
Facial no data
Vascular no data
16. Technology - Limitations
Biometric Record Data Size (Bytes)
Retinal Scan 96
Iris Scan 512
Fingerprints 512-1000
Hand Geometry 9
Signature 3900
Voice 60/word
Facial Recognition 100-3500
17. Technology - Limitations
Verification versus Identification
•User base of millions
•Real-time verification precluded
•Token with template possible
•Verify identity against template
18. Technology - Limitations
Transaction/User Base Examples
• In May 2000 there were
over 400,000 transactions at
Exxon and Mobil C-stores.
• For 2002 projects to over
35 million transactions.
• 185,000 ATMs in the US
• Close to 15B transactions in
US in 2002 (declining)
19. Technology - Limitations
Templates - Storage & Processing
•Local to Local - within a smart card
•Local and Terminal - template on
smartcard copied to terminal to be verified.
•Remote and Terminal - template stored
remotely and compared to local template
•Remote and Remote - locally acquired
template is compared on remote system to remote
template.
20. Technology - Limitations
• Failure To Enroll - critical to get good measurement at this stage.
Some will choose not to use the system and must still be processed.
• Failure To Acquire - systems that require high level of user
cooperation.
• False Positive (FAR) - users must trust and accept
• False Negative (FRR) - how many rejected falsely?
• EER (Equal Error Rate Crossover)
21. Technology - Limitations
• False Acceptance Rate (FAR): The chance that an imposter will
be recognized (obtain a higher score) at a certain threshold.
• False Rejection Rate (FRR): The chance that the correct
person will not obtain a score above a certain threshold.
• Both the FAR and FRR are functions of threshold. The
threshold where the two probabilities are the same is the
Equal Error Rate (EER). For example, if the EER is 1%, that
means 1% is the right people are rejected and 1% of the
wrong people are accepted above a certain threshold.
22. Technology - Limitations
• Crossover point between FAR and FRR known as
the Equal Error rate (EER) and this is often
quoted. Argued that the lower this figure is, the
better the system performance is. Unfortunately,
how the figure was established can seriously
affect it. A system that performs well in the
laboratory with trained, co-operative users will
generate a completely different set of values
with inexperienced or less co-operative users.
23. Technology - Limitations
Probability
Distribution
EER
Genuine Imposter
Score
(e.g. hamming distance)
false accept false reject
24. Technology - Limitations
Centre for Mathematics and Scientific Computing National Physical Laboratory Middlesex, UK
“Biometric Product Testing Final Report”
Issue 1.0 March 19, 2001
26. Technology - Limitations
The Most Intriguing - The Iris
Forms during 3rd month of
gestation, 8th month Pattern
complete, coloration through crypts
Iris radial furrows
after birth
Only internal organ of the body
that is normally visible externally pigment frill
Pupil
(highly protected
by cornea and eyelid) pupilary area
Impossibility of surgically ciliary area
modifying it without
Sclera
unacceptable risk to vision collarette
27. Technology - Limitations
The Most Intriguing - The Iris
• Iris verification has been, and still is, used in ATMs
(Bank United in Texas for example). The size and the
cost (>$1000 for an ATM compatible system) seem to
be the main reasons for this system not growing
further in this area. The risk of fraud by the use of
force is possible. The thought of attempting fraud by
mutilation of a user is distasteful but could exist.
28. Technology - Limitations
Other Considerations
• Enrolling Users
• Sheer number of Users
• How invasive?
• How stored and available how?
• Still PIN systems, will users forget?
29. Facial Recognition
• Humans Are Easily Fooled By
Pictures
• Vendors: Visionics, Bio4, Viisage
(Lau), eTrue, Imagis…
• Pros; Least Invasive, Fast
• Issues: lighting, aging… general
error rates
• Existing Database – Usability?
30. Facial Recognition
Methods for Matching
• Automatic face processing.
• Neural Network processing
• Eigenfaces
• Local Feature Analysis.
31. Facial Recognition
• Facial Recognition Vendor Test 2000
(FRVT 2000)
- Sponsors: Counterdrug Tech Prog Office, NIST, DARPA
- Performance Update - State of technology & Specific
Vendors
- Update of FERET 1993-1998
(FacEREcognitionTechnology)
- 13,872 Images of 1462 subjects In nxn Test
(192 Million Comparisons over 72 hrs)
- Results Summarized in 57 Figures
32. Facial Recognition
FRVT 2000 Test Overview
Banque-
Test Visionics Lau C-VIS Miros Tec Bio4
Expression 1 2 3 - - -
Illumination 1 2 3 - - -
Pose 1 2 3 - - -
Media 1 2 3 - - -
Distance 2 1 3 - - -
Temporal 1 2 3 - - -
Resolution 1 2 3 - - -
1-Best, 3-Worst
General Observations
Facial May Not Be Suitable For Exact “Identification”
(needs human)
Best Case Verification EER ~.02
Distribution Not Symmetrical, Use ROC Not EER
34. Facial Accuracy
Processing False Accepts
• Premise: Load Terrorist Picture Database Into System
• Issues: Quality of enrollment pictures
• Impact Assessment
– e.g. Hartfield Airport
• 150M People Year -> ~411,000 / Day
• 2400 Flights / Day
– @ An Optimistic EER of 2%
• 8219 False Accepts
• @ 20 minutes / False Accept -> 2739 manhours / day
• P(8 false accepts | 375 passengers) = 8.6%
35. Facial Vendors
• Visionics
• Miros
• Bio4
• Viisage
• There are many others...
36. Facial Vendors
Visionics
• Fundamental to any face recognition system is the way
in which faces are coded. Visionics FaceIt® uses Local
Feature Analysis (LFA) to represent facial images in
terms of local statistically derived building blocks.
• LFA is a mathematical technique developed by co-
founders of Visionics Corporation, and is based on the
realization that all facial images (for that matter all
complex patterns) can be synthesized from an irreducible
set of building elements.
37. Facial Vendors
Visionics
• They span multiple pixels (but are still local) and
represent universal facial shapes, but are not exactly the
commonly known facial features. In fact, there are many
more facial building elements than there are facial parts.
However, it turns out that synthesizing a given facial
image, to a high degree of precision, requires only a
small subset (12-40 characteristic elements) of the total
available set. Identity is determined not only by which
elements are characteristic, but also by the manner in
which they are geometrically combined (i.e. their relative
positions).
38. Applications & Advantages
Authentication
• Computer/Network Security
• Banking
• Smart Cards
• Access Control
• Border Control
40. Applications & Advantages
Human ID at a Distance
• Surveillance
• CCTV
• Human Traffic Control
• Friend or Foe
41. Applications & Advantages
ID Solutions
• Voter Registration
• National Ids
• Passports
• Drivers Licenses
• Employee IDs
42. Applications & Advantages
ID Examples
• Eliminating Aliases & Duplicates
• Mexican Election System
• West Virginia
• Colorado State DMV (under
installation)
• Dominican Republic
• U.S. State Department Visa
Issuance Program
43. Applications & Advantages
• UK Prime Minister Tony
Blair visits the FaceIt®
Surveillance installation
in Newham, where
crime has been reduced
by 34% overall.
44. Applications & Advantages
• ATMs
• Integrating facial biometrics systems into an
ATM is a real possibility as in many cases the
sensing system (camera) is already there and
could easily be used.
• Persons trying fraud via decapitation of the
legitimate user is gruesome but imaginable.
The use of masks is another avenue for
potential fraud
45. Applications & Advantages
• Facial recognition technology is the only
commercially-available biometric capable
of identifying humans at a distance. It
has already been deployed in some high-
profile locations -- in casinos, European
soccer matches and in town centers --
and has shown significant results.
46. Applications & Advantages
January 14: British Virgin Islands Select
AiT's Automated Border Management
System: $1.1 million contract for enTReX
• At each inspection point, AiT's imPAX Reader will be used
to capture images of travellers ID documents, and digital
cameras will capture live facial images. All of the images
and data captured will be sent to enTReX, a software-
based border management system, for processing. The
system provides better tools to deal with undocumented
travellers, identify those without a legitimate right to be
on the islands, and detect mismatches between those
exiting the border and those who entered.
47. Applications & Advantages
January 4, 2002 CNET
Visionics signed a deal late last year with
conglomerate Tyco International to distribute
its technology at some 100 of the nation's 450
commercial airports. Even the U.S. Army
recently licensed face-recognition technology
from rival Viisage Technology to create custom
high-security applications.
48. Is Visionics the Future?
An industry in search of a compelling mass
application
No clearly appreciated value proposition
(convenience can only go so far)
Several adoption barriers
Privacy concerns
Political will to change
Lack of infrastructure & funding
49. Is Visionics the Future?
Positives
ID technologies: a corner stone of
defense against terrorism & crime
Safety & Security: clear and present
value propositions
Accelerated funding & federal security
mandates
Favorable Public Opinion
50. Is Visionics the Future?
Positives
• Security is no longer viewed as a drag on bottom
line: at least for now.
• WTC disaster will cost NYC an $100 billion.
Investing in security is like buying insurance
• Significant funding for security programs —
federal, state, local & commercial:
– AIP : Aviation Trust Fund is up
– PFC : increased by up to $5 per one way trip
51. Is Visionics the Future?
New Drivers' Licenses Study Underway
• WASHINGTON (AP) - January 8, 2002 --
The government is working with the
states to develop a new generation of
drivers' licenses that could be checked
anywhere and would contain electronically
stored information such as fingerprints for
the country's 184 million licensed drivers.
52. Is Visionics the Future?
Camera May Be Able to Spot Liars
• Minnesota - January 3, 2002 -- A heat-
sensing camera trained on people's faces was able
to detect liars in a study. In six of eight people who
lied, the high-resolution thermal imaging camera
detected a faint blushing around their eyes that
Mayo Clinic researchers said is evidence of
deception. Such facial imaging, they said, could
provide a simple and rapid way of scanning people
being questioned at airports or border crossings.
53. Is Visionics the Future?
Net Nanny Signs International Distributors
For BioPassword
• December 4, 2002 -- Net Nanny Software
International Inc. (OTCBB:NNSWF - news;
CDNX:NNS.V) announced today that it has signed
Wildspace, Ltd., Junek Ltd. and Joint Future
Systems, S.C., which are headquartered in the UK,
Czech Republic and Mexico respectively, to
distribute its strong user authentication
technology, BioPassword®.
54. Is Visionics the Future?
Net Nanny Signs International Distributors
For BioPassword
• All three companies will distribute BioPassword, the
authentication solution for enterprise networks,
and the BioPassword software developer's kit,
which enables third parties to incorporate the
patented keystroke dynamics technology into their
own applications.
56. Is Visionics the Future?
Issues To Deal With
• Privacy -- for verification rather
than identification the issue is
smaller. Orwellian perception.
57. Is Visionics the Future?
ACLU in January 2002 noted:
• "Face recognition is all hype and no action," Barry
Steinhardt, associate director of the ACLU, said in a
statement. "Potentially powerful surveillance systems
like face recognition need to be examined closely
before they are deployed, and the first question to ask
is whether the system will actually improve our safety.
The experience of the Tampa Police Department
confirms that this technology doesn't deliver."
58. Is Visionics the Future?
Conclusions
• Commodity level implementation still 2 to 3
years away.
• Government is the big driver.
• Smartcards, ID Cards, Mobile, Credit Card
are some of the new Driver Enablers
59. Visionics
Personal Thanks and Appreciation for
the Encouragement provided by these
leading companies:
• Symbol Technologies
• NCR Corporation
• Visionics
• Wincor Nixdorf
60. Visionics
Thank You
For More Information
• Visit Kiosks.org Association at www.kiosks.org/smi
• The UK Biometrics Working Group, Biometric Product Testing, Final Report. At
http://www.cesg.gov.uk/technology/biometrics/media/Biometric%20Test%20Report%20pt1.pdf
• Biometric Market Report 2000-2005,
http://www.biometricgroup.com/e/biometric_market_report.htm
• Visionics Corporation http://www.visionics.com/faceit/apps/auth.html
• Neusciences - Biometrics http://www.neusciences.com/Biometrics/applications.htm
• Miros http://www.miros.com/
• Advanced Biometrics Inc. http://www.livegrip.com/
• Iridian http://www.iridiantech.com/
• The Defense Advanced Research Projects Agency (DARPA) Human ID at a Distance
http://www.darpa.mil/ito/research/hid/
61. Visionics - Addendum
Craig Keefner is the Executive Director of Kiosks.org Association. The
Association is worldwide and includes both vendors and users/deployers. It’s
mission is to identify and promote the interests of companies engaged in the
electronic, self-service kiosk industry.
Working committees include: Marketing and Public Relations, Best Practices,
Technology and Standards, Government Relations, Research and Statistics,
Advertising and Developing Markets. The President of the Association is Mr..
Richard Rommel of Eastman Kodak and the Chairman is Mr.. Richard Good of
NetWorld Alliance.
Members include: NCR Corporation, Symbol Technologies, Wincor-Nixdorf, IBM,
Compaq, Intel, Eastman Kodak, MEI, Netshift, Eurocoin, ePOINT and many
many others.
For more information or to join: go online at www.kiosks.org/join or call 1-866-
240-1318. You can email us at info@kiosks.org
This presentation delivered February 6, 2002 at The Hatton for Kiosks 2002