©Veridium All Rights Reserved
Top Biometric Identifiers:
Risks & Rewards
B E FORE W E B E G IN
Attendees have been muted
You may submit questions at any
time, but we will respond at the
conclusion of the presentation
during the Q&A session
©Veridium All Rights Reserved
Asem Othman, PhD
Biometric Scientist
• PhD in Electrical Engineering from West Virginia University.
Post-Doc Fellow at Michigan State University
• Holds 3 patents/patents pending related to biometrics
• Lead scientist behind our 4 Fingers TouchlessID technology &
Distributed Data Model using Visual Cryptography
BE F O RE WE BE G IN
©Veridium All Rights Reserved
• The Power of Mobile Biometrics
• Top Five Biometric Identifiers
• The Privacy Issue
• How We Address Privacy Concerns
AGENDA
©Veridium All Rights Reserved
P ROV IN G ID E N T IT Y
©Veridium All Rights Reserved
Passwords only authorize
access, while
authentication is the
process of verifying the
claim that identity makes.
Identity is the “claim” you
make to access information,
but making that claim with a
password doesn’t prove who
you are.
Only biometrics – your
face, your voice, your
fingerprints – prove
you are who you say
you are.
Universality (Does every user have it?)
Distinctiveness (Is it unique across users?)
Permanence (Does it change over time?)
Collectability (Can it be measured quantitatively?)
H OW TO CH O O SE A BIO ME T RIC T RAIT
©Veridium All Rights Reserved
A biometric recognition system is a pattern recognition system
that recognize individuals based on their biometric trait(s)
BIO ME T RIC SYST E MS
©Veridium All Rights Reserved
• Touch ID made it easy (and
convenient) for the consumer
to understand biometrics.
• Accessing your phone
• Apple Pay
TO UCH ID
©Veridium All Rights Reserved
Finger
Face
Iris
Voice
Traditional Hardware Veridium Technologies
All major biometrics can be captured using different maneuvers with a smartphone
O UR AP P ROACH TO BIO ME T RICS
©Veridium All Rights Reserved
• Mobile devices have changed
the fundamental way we
interact with technology.
• Mobile devices allow for the
elimination of proprietary
hardware, moving beyond
traditional biometric scanners.
MO BILE BIO ME T RICS
©Veridium All Rights Reserved
Performance (Does it meet error rate, throughput..?)
Users Perception and Acceptability (Is it acceptable to users?)
Vulnerability (Can it be easily spoofed?)
App Integration (Can it be acquired by available devices?)
Large-Scale Adoption ”Usability” (What is the post-usage
attitude?)
“A P R AC T I C A L M O B I L E B I O M E T R I C ”
H OW TO CH O O SE A BIO ME T RIC T RAIT
©Veridium All Rights Reserved
TOP BI OMET R I C S
FACE
• People are used to taking selfies.
• It’s socially accepted.
• It’s well established.
• Reliable face recognition systems are already in use.
FACE AS A MO BILE BIO ME T RIC
©Veridium All Rights Reserved
• Face – We Naturally Use It
• The location and shape of facial
attributes.
• The overall analysis of the face.
• In practice, a facial recognition system
should automatically …
• Detect the face.
• Locate the facial attribute.
• Recognize the face.
Shape
Color
Texture
Requiring a simple background and illumination
FACE AS A MO BILE BIO ME T RIC
©Veridium All Rights Reserved
The database is designed to test difficult real world situations that a face system must cope with.
The leading matcher algorithms performed very similarly on our test databases.
~3% FRR at 1% FAR for controlled illumination and office environments.
30-50% FRR at 1% FAR for whole database.
V E RID IUM FACE DATABASE
©Veridium All Rights Reserved
• Post-usage attitude:
• Pose variation
• Taking selfie in the middle of a
meeting, indoor, or outdoor
• Older generation
• Culture
• Womens‘ appearance concerns
• Twins and look-alikes
• Easily spoofed
ISSUE S
©Veridium All Rights Reserved
VOICE
• Voice is a combination of physiological and
behavioral biometrics.
• An individual’s voice is based on the shape and size
of the appendages (e.g., vocal tracts, mouth, nasal
cavities, and lips) that are used in the synthesis of
the sound.
• Natural signal to produce.
• No visual contact is required.
• No special equipment is required.
• Can be done while doing other things.
VO ICE AS A BIO ME T RIC
©Veridium All Rights Reserved
• Text-dependent recognition
– Recognition system knows text was spoken by a person
– Examples: Fixed phrase, prompted phrase
– Used for applications with strong control over user input
– Knowledge of spoken text can improve system performance
• Text-independent recognition
– Recognition system does not know text was spoken by a
person
– Examples: User selected phrase, conversational speech
– Used for applications with less control over user input
– More flexible system but also more difficult problem
– Speech recognition can provide knowledge of spoken text
SP E E CH MO DALIT IE S
©Veridium All Rights Reserved
• The physiological characteristics of human speech are unvaried for an individual, but the
behavioral part of the speech of a person changes over time due to age, medical
conditions (such as common cold), emotional state, etc.
• Voice is also not very distinctive and may not be appropriate for large-scale identification.
• Acoustic environment and background noise
• Presentation attacks
• Tape recordings
• Identical twins / sound-alikes
• A disruption to working memory*
*Trewin, Shari, et al. "Biometric authentication on a mobile device: a study of user effort, error and task disruption." Proceedings of the 28th Annual Computer
Security Applications Conference. ACM, 2012.
ISSUE S
©Veridium All Rights Reserved
FINGERPRINT
• Is the most well recognized biometric.
• A fingerprint is the pattern of ridges and valleys on
the surface of a fingertip.
• The formation of the fingerprint is determined
during the first seven months of fetal
development.
• Fingerprints of identical twins are different, as are
the prints on each finger of the same person.
• Fingerprints don’t change over time.
• It’s widely believed fingerprints are unique.
F IN G E RP RIN T AS A BIO ME T RIC
©Veridium All Rights Reserved
SMART P H O N E S WIT H F IN G E RP RIN T SE N SO RS
©Veridium All Rights Reserved
Ridge ending Bifurcation Short ridge
Minutia
Type x y θ
Ridge 35 24 12’
... ... ... ...
... ... ... ...
H OW F IN G E RP RIN T MATCH IN G WO RK S
©Veridium All Rights Reserved
• Proprietary acquisition sensors
• Scars and wet fingers
• Presentation attack
• Spoofing
ISSUE S
©Veridium All Rights Reserved
IRIS
• The only internal “protected” organ readily
visible from the outside of a person.
• Unlike fingerprints, capturing the iris biometric
is similar to taking a picture.
• The iris pattern has a high degree of
randomness.
• Different even for identical twins.
• The pattern is stable through lifetime .
• Extremely difficult to surgically tamper the
texture of the iris.
IRIS AS A MO BILE BIO ME T RIC
©Veridium All Rights Reserved
Light Irises
“LIK E TAK IN G A P ICT URE ”
©Veridium All Rights Reserved
SMART P H O N E S WIT H IRIS SE N SO RS
©Veridium All Rights Reserved
©Veridium All Rights Reserved
• Proprietary acquisition sensor.
• Small target to acquire from a distance.
• Located behind a curved, wet, reflecting surface.
• Obscured by eyelashes,lenses.
• Partially occluded by eyelids, often drooping.
• Deforms non-elastically as pupil changes size.
• Uncomfortable for some.
Deviated gaze
Accessories
Illumination
Deformations
Motion blur
Occlusions
ISSUE S
©Veridium All Rights Reserved
4 FINGERS
TouchlessID
• Multimodal Biometric system
• Consolidate the evidence presented by
multiple biometric sources.
• Typically provides better recognition
performance compared to systems
based on a single biometric modality.
• Provides anti-spoofing measures by
making it difficult for an intruder to
spoof multiple biometric instance
simultaneously.
Patent US 9,361,507
©Veridium All Rights Reserved
©Veridium All Rights Reserved
• False rejection rate (FRR) is as low as 1% at false acceptance rate (FAR) of 0.01%.
• 4 Fingers has its own light source (your phone’s flash) so it works in any lighting
condition
• 4 Fingers requires no external hardware.
• We just require a 5MP camera and LED flash.
• 4 Fingers is one of the most secure biometrics available.
©Veridium All Rights Reserved
4 Fingers is reliable in almost any environment
4 Fingers captured print 4 Fingers processed print Standard Fingerprint
Sensor
Minutiae
The contactless and contact prints are of comparable quality.
The minutia match.
©Veridium All Rights Reserved
Print Quality
• The Cooperative Research and Development Agreement (CRADA)
allows NIST to work with U.S. industry, academia and other
organizations on cooperative R&D projects
• Build upon NIST's research in developing methodologies for
measuring the image fidelity of contactless fingerprint capture
devices
• Produce open testing methods, metrics, and artifacts that will
support future certification of contactless fingerprint devices for
inclusion on Government Certified Products Lists
©Veridium All Rights Reserved
Contactless Fingerprint Capture Device
Measurement Research Program
• Missing finger
• Gloves
• Requires both hands
ISSUE S
©Veridium All Rights Reserved
COMPARISON
CO MPARISO N O F TO P “MO BILE ” BIO ME T RIC T E CH N O LO GIE S
©Veridium All Rights Reserved
(H=High, M=Medium, L=Low)
Biometrics Universality Uniqueness Permanence Collectability Performance Acceptability
Face H L M H L H
Fingerprint M H H M* H M
Iris H H H M* H L*
Voice M L L M* L H
4 Fingers
TouchlessID*
H H H H H H
(H=High, M=Medium, L=Low)
Circumventions
(Presentation
Attack)
H
M
L
H
L
Anil K. Jain, Arun Ross, and Salil Prabhakar. "An introduction to biometric recognition." Appeared in IEEE Transactions on Circuits and Systems for Video
Technology, Special Issue on Image- and Video-Based Biometrics, Vol. 14, No. 1, January 2004.
• Familiarity with technology
• Experience with device
• Environment of use (e.g., indoor/outdoor, home/work, car/bus,…)
• Transaction value (e.g., small, medium, or large transactions)
• Time consuming tasks (e.g., enrollment, etc.)
USABILIT Y ISSUE S AN D LARG E - SCALE AD O P T IO N IN BIO ME T RICS
©Veridium All Rights Reserved
• Each biometric has its strengths and weaknesses.
• Liveness, Presentation Attack, Spoofing.
• No biometric is “optimal,” but many are “admissible.”
F IN AL WO RD S
©Veridium All Rights Reserved
PRIVACY
AND
SECURITY RISKS
• Biometric data of an individual is often stored in a central database
• Raises issues related to security and privacy of biometric data
• Unlike compromised passwords, it is difficult tore-issue biometric data
• Cross-database matching may be done to track individuals
P RE SE RV IN G DATA P RIVACY
©Veridium All Rights Reserved
O UR ST RAT E G Y – D IST RIBUT E D DATA
©Veridium All Rights Reserved
MATCH IN G – MO BILE /SE RV E R
©Veridium All Rights Reserved
Cryptography is simply the art of sending and receiving coded messages.
CRYP TO G RAP H Y
©Veridium All Rights Reserved
• Visual Cryptography Scheme (VCS) is a simple and secure way to allow the
secret sharing of secrets without any cryptographic computations.
• It is the encryption of visual information such that decryption can be
performed using the human visual system.
• Someone who has no previous knowledge of Cryptography.
• The mathematical proof of this scheme and its perfect encryption are shown in
the original paper by Naor & Shamir.
* M. Naor and A. Shamir, “Visual cryptography,” in EUROCRYPT, pp. 1–12, 1994.
V ISUAL CRYP TO G RAP H Y
©Veridium All Rights Reserved
V ISUAL CRYP TO G RAP H Y
©Veridium All Rights Reserved
For more information contact: Info@VeridiumID.com
Phone: +1 212.231.0011 • www.VeridiumID.com • Twitter: @VeridiumID • LinkedIn: Veridium
QUESTIONS?
©Veridium All Rights Reserved

Top Biometric Identifiers: Risks & Rewards

  • 1.
    ©Veridium All RightsReserved Top Biometric Identifiers: Risks & Rewards
  • 2.
    B E FOREW E B E G IN Attendees have been muted You may submit questions at any time, but we will respond at the conclusion of the presentation during the Q&A session ©Veridium All Rights Reserved
  • 3.
    Asem Othman, PhD BiometricScientist • PhD in Electrical Engineering from West Virginia University. Post-Doc Fellow at Michigan State University • Holds 3 patents/patents pending related to biometrics • Lead scientist behind our 4 Fingers TouchlessID technology & Distributed Data Model using Visual Cryptography BE F O RE WE BE G IN ©Veridium All Rights Reserved
  • 4.
    • The Powerof Mobile Biometrics • Top Five Biometric Identifiers • The Privacy Issue • How We Address Privacy Concerns AGENDA ©Veridium All Rights Reserved
  • 5.
    P ROV ING ID E N T IT Y ©Veridium All Rights Reserved Passwords only authorize access, while authentication is the process of verifying the claim that identity makes. Identity is the “claim” you make to access information, but making that claim with a password doesn’t prove who you are. Only biometrics – your face, your voice, your fingerprints – prove you are who you say you are.
  • 6.
    Universality (Does everyuser have it?) Distinctiveness (Is it unique across users?) Permanence (Does it change over time?) Collectability (Can it be measured quantitatively?) H OW TO CH O O SE A BIO ME T RIC T RAIT ©Veridium All Rights Reserved
  • 7.
    A biometric recognitionsystem is a pattern recognition system that recognize individuals based on their biometric trait(s) BIO ME T RIC SYST E MS ©Veridium All Rights Reserved
  • 8.
    • Touch IDmade it easy (and convenient) for the consumer to understand biometrics. • Accessing your phone • Apple Pay TO UCH ID ©Veridium All Rights Reserved
  • 9.
    Finger Face Iris Voice Traditional Hardware VeridiumTechnologies All major biometrics can be captured using different maneuvers with a smartphone O UR AP P ROACH TO BIO ME T RICS ©Veridium All Rights Reserved
  • 10.
    • Mobile deviceshave changed the fundamental way we interact with technology. • Mobile devices allow for the elimination of proprietary hardware, moving beyond traditional biometric scanners. MO BILE BIO ME T RICS ©Veridium All Rights Reserved
  • 11.
    Performance (Does itmeet error rate, throughput..?) Users Perception and Acceptability (Is it acceptable to users?) Vulnerability (Can it be easily spoofed?) App Integration (Can it be acquired by available devices?) Large-Scale Adoption ”Usability” (What is the post-usage attitude?) “A P R AC T I C A L M O B I L E B I O M E T R I C ” H OW TO CH O O SE A BIO ME T RIC T RAIT ©Veridium All Rights Reserved
  • 12.
    TOP BI OMETR I C S
  • 13.
  • 14.
    • People areused to taking selfies. • It’s socially accepted. • It’s well established. • Reliable face recognition systems are already in use. FACE AS A MO BILE BIO ME T RIC ©Veridium All Rights Reserved
  • 15.
    • Face –We Naturally Use It • The location and shape of facial attributes. • The overall analysis of the face. • In practice, a facial recognition system should automatically … • Detect the face. • Locate the facial attribute. • Recognize the face. Shape Color Texture Requiring a simple background and illumination FACE AS A MO BILE BIO ME T RIC ©Veridium All Rights Reserved
  • 16.
    The database isdesigned to test difficult real world situations that a face system must cope with. The leading matcher algorithms performed very similarly on our test databases. ~3% FRR at 1% FAR for controlled illumination and office environments. 30-50% FRR at 1% FAR for whole database. V E RID IUM FACE DATABASE ©Veridium All Rights Reserved
  • 17.
    • Post-usage attitude: •Pose variation • Taking selfie in the middle of a meeting, indoor, or outdoor • Older generation • Culture • Womens‘ appearance concerns • Twins and look-alikes • Easily spoofed ISSUE S ©Veridium All Rights Reserved
  • 18.
  • 19.
    • Voice isa combination of physiological and behavioral biometrics. • An individual’s voice is based on the shape and size of the appendages (e.g., vocal tracts, mouth, nasal cavities, and lips) that are used in the synthesis of the sound. • Natural signal to produce. • No visual contact is required. • No special equipment is required. • Can be done while doing other things. VO ICE AS A BIO ME T RIC ©Veridium All Rights Reserved
  • 20.
    • Text-dependent recognition –Recognition system knows text was spoken by a person – Examples: Fixed phrase, prompted phrase – Used for applications with strong control over user input – Knowledge of spoken text can improve system performance • Text-independent recognition – Recognition system does not know text was spoken by a person – Examples: User selected phrase, conversational speech – Used for applications with less control over user input – More flexible system but also more difficult problem – Speech recognition can provide knowledge of spoken text SP E E CH MO DALIT IE S ©Veridium All Rights Reserved
  • 21.
    • The physiologicalcharacteristics of human speech are unvaried for an individual, but the behavioral part of the speech of a person changes over time due to age, medical conditions (such as common cold), emotional state, etc. • Voice is also not very distinctive and may not be appropriate for large-scale identification. • Acoustic environment and background noise • Presentation attacks • Tape recordings • Identical twins / sound-alikes • A disruption to working memory* *Trewin, Shari, et al. "Biometric authentication on a mobile device: a study of user effort, error and task disruption." Proceedings of the 28th Annual Computer Security Applications Conference. ACM, 2012. ISSUE S ©Veridium All Rights Reserved
  • 22.
  • 23.
    • Is themost well recognized biometric. • A fingerprint is the pattern of ridges and valleys on the surface of a fingertip. • The formation of the fingerprint is determined during the first seven months of fetal development. • Fingerprints of identical twins are different, as are the prints on each finger of the same person. • Fingerprints don’t change over time. • It’s widely believed fingerprints are unique. F IN G E RP RIN T AS A BIO ME T RIC ©Veridium All Rights Reserved
  • 24.
    SMART P HO N E S WIT H F IN G E RP RIN T SE N SO RS ©Veridium All Rights Reserved
  • 25.
    Ridge ending BifurcationShort ridge Minutia Type x y θ Ridge 35 24 12’ ... ... ... ... ... ... ... ... H OW F IN G E RP RIN T MATCH IN G WO RK S ©Veridium All Rights Reserved
  • 26.
    • Proprietary acquisitionsensors • Scars and wet fingers • Presentation attack • Spoofing ISSUE S ©Veridium All Rights Reserved
  • 27.
  • 28.
    • The onlyinternal “protected” organ readily visible from the outside of a person. • Unlike fingerprints, capturing the iris biometric is similar to taking a picture. • The iris pattern has a high degree of randomness. • Different even for identical twins. • The pattern is stable through lifetime . • Extremely difficult to surgically tamper the texture of the iris. IRIS AS A MO BILE BIO ME T RIC ©Veridium All Rights Reserved
  • 29.
    Light Irises “LIK ETAK IN G A P ICT URE ” ©Veridium All Rights Reserved
  • 30.
    SMART P HO N E S WIT H IRIS SE N SO RS ©Veridium All Rights Reserved
  • 31.
  • 32.
    • Proprietary acquisitionsensor. • Small target to acquire from a distance. • Located behind a curved, wet, reflecting surface. • Obscured by eyelashes,lenses. • Partially occluded by eyelids, often drooping. • Deforms non-elastically as pupil changes size. • Uncomfortable for some. Deviated gaze Accessories Illumination Deformations Motion blur Occlusions ISSUE S ©Veridium All Rights Reserved
  • 33.
  • 34.
    • Multimodal Biometricsystem • Consolidate the evidence presented by multiple biometric sources. • Typically provides better recognition performance compared to systems based on a single biometric modality. • Provides anti-spoofing measures by making it difficult for an intruder to spoof multiple biometric instance simultaneously. Patent US 9,361,507 ©Veridium All Rights Reserved
  • 35.
  • 36.
    • False rejectionrate (FRR) is as low as 1% at false acceptance rate (FAR) of 0.01%. • 4 Fingers has its own light source (your phone’s flash) so it works in any lighting condition • 4 Fingers requires no external hardware. • We just require a 5MP camera and LED flash. • 4 Fingers is one of the most secure biometrics available. ©Veridium All Rights Reserved 4 Fingers is reliable in almost any environment
  • 37.
    4 Fingers capturedprint 4 Fingers processed print Standard Fingerprint Sensor Minutiae The contactless and contact prints are of comparable quality. The minutia match. ©Veridium All Rights Reserved Print Quality
  • 38.
    • The CooperativeResearch and Development Agreement (CRADA) allows NIST to work with U.S. industry, academia and other organizations on cooperative R&D projects • Build upon NIST's research in developing methodologies for measuring the image fidelity of contactless fingerprint capture devices • Produce open testing methods, metrics, and artifacts that will support future certification of contactless fingerprint devices for inclusion on Government Certified Products Lists ©Veridium All Rights Reserved Contactless Fingerprint Capture Device Measurement Research Program
  • 39.
    • Missing finger •Gloves • Requires both hands ISSUE S ©Veridium All Rights Reserved
  • 40.
  • 41.
    CO MPARISO NO F TO P “MO BILE ” BIO ME T RIC T E CH N O LO GIE S ©Veridium All Rights Reserved (H=High, M=Medium, L=Low) Biometrics Universality Uniqueness Permanence Collectability Performance Acceptability Face H L M H L H Fingerprint M H H M* H M Iris H H H M* H L* Voice M L L M* L H 4 Fingers TouchlessID* H H H H H H (H=High, M=Medium, L=Low) Circumventions (Presentation Attack) H M L H L Anil K. Jain, Arun Ross, and Salil Prabhakar. "An introduction to biometric recognition." Appeared in IEEE Transactions on Circuits and Systems for Video Technology, Special Issue on Image- and Video-Based Biometrics, Vol. 14, No. 1, January 2004.
  • 42.
    • Familiarity withtechnology • Experience with device • Environment of use (e.g., indoor/outdoor, home/work, car/bus,…) • Transaction value (e.g., small, medium, or large transactions) • Time consuming tasks (e.g., enrollment, etc.) USABILIT Y ISSUE S AN D LARG E - SCALE AD O P T IO N IN BIO ME T RICS ©Veridium All Rights Reserved
  • 43.
    • Each biometrichas its strengths and weaknesses. • Liveness, Presentation Attack, Spoofing. • No biometric is “optimal,” but many are “admissible.” F IN AL WO RD S ©Veridium All Rights Reserved
  • 44.
  • 45.
    • Biometric dataof an individual is often stored in a central database • Raises issues related to security and privacy of biometric data • Unlike compromised passwords, it is difficult tore-issue biometric data • Cross-database matching may be done to track individuals P RE SE RV IN G DATA P RIVACY ©Veridium All Rights Reserved
  • 46.
    O UR STRAT E G Y – D IST RIBUT E D DATA ©Veridium All Rights Reserved
  • 47.
    MATCH IN G– MO BILE /SE RV E R ©Veridium All Rights Reserved
  • 48.
    Cryptography is simplythe art of sending and receiving coded messages. CRYP TO G RAP H Y ©Veridium All Rights Reserved
  • 49.
    • Visual CryptographyScheme (VCS) is a simple and secure way to allow the secret sharing of secrets without any cryptographic computations. • It is the encryption of visual information such that decryption can be performed using the human visual system. • Someone who has no previous knowledge of Cryptography. • The mathematical proof of this scheme and its perfect encryption are shown in the original paper by Naor & Shamir. * M. Naor and A. Shamir, “Visual cryptography,” in EUROCRYPT, pp. 1–12, 1994. V ISUAL CRYP TO G RAP H Y ©Veridium All Rights Reserved
  • 51.
    V ISUAL CRYPTO G RAP H Y ©Veridium All Rights Reserved
  • 52.
    For more informationcontact: Info@VeridiumID.com Phone: +1 212.231.0011 • www.VeridiumID.com • Twitter: @VeridiumID • LinkedIn: Veridium QUESTIONS? ©Veridium All Rights Reserved