Biometrics Biometrics and and Authentication Authentication

spotlessstareΑσφάλεια

29 Νοε 2013 (πριν από 3 χρόνια και 7 μήνες)

55 εμφανίσεις

Biometrics
Biometrics
,
,
Authentication
Authentication
©
©
2008 M
2008 M
ü
ü
ller
ller
Berne University of
Berne University of
Applied Sciences
Applied Sciences
Biometrics
Biometrics
and
and
Authentication
Authentication
Biometric methods
Overview, notions, classification, quality
parameters, systems, threads, methods
Biometrics
Biometrics
,
,
Authentication
Authentication
2
2
©
©
2008 M
2008 M
ü
ü
ller
ller
Berne University of
Berne University of
Applied Sciences
Applied Sciences
Inhalt
Inhalt
Biometric notions
Biometric system architecture
Biometric processing
Implementation
Performance and quality parameters
Comparison curves
Advantages, threads and problems
Requirements
Biometric methods
Economical impact
Biometrics
Biometrics
,
,
Authentication
Authentication
3
3
©
©
2008 M
2008 M
ü
ü
ller
ller
Berne University of
Berne University of
Applied Sciences
Applied Sciences
Basic
Basic
notions
notions
Initial registration of biometric data of a person, creation of a
reference template
Enrolment
Capture of the biometric characteristics of a person with unknown
identity, processing and comparison of the captured template with all
reference templates of the enrolled persons (1:n)
Identification
Capture of the biometric characteristics of a person with an assumed
(claimed) identity, processing and comparison of the captured
template with the reference template of the assumed persons (1:1)
Verification
General notion for the identity examination using verification or
identification processes (not only biometrics). In biometrics the
general notion of such an examination is biometric identity
recognition.
Authentication
JTC1/SC37 (ISO):
“The automated recognition of individuals based on their
behavioral and biological characteristics”
Biometrics
Biometrics
Biometrics
,
,
Authentication
Authentication
4
4
©
©
2008 M
2008 M
ü
ü
ller
ller
Berne University of
Berne University of
Applied Sciences
Applied Sciences
Basic
Basic
system
system
architecture
architecture
Biometric System
Identity
Information
Identity
Information
Individual
Captured
Biometric
sample
Biometric
comparison
Identity
information
Identification
Identity
Credential
Verification
Identity
Database
T
e
m
p
l
a
t
e
a
s
s
i
g
n
m
e
n
t
Identity for
reference
template
Reference
template
storage
Matching
decision
Sen-
sor
Biometrics
Biometrics
,
,
Authentication
Authentication
5
5
©
©
2008 M
2008 M
ü
ü
ller
ller
Berne University of
Berne University of
Applied Sciences
Applied Sciences
Components
Components
of a
of a
biometric
biometric
system
system
Biometrics
Biometrics
,
,
Authentication
Authentication
6
6
©
©
2008 M
2008 M
ü
ü
ller
ller
Berne University of
Berne University of
Applied Sciences
Applied Sciences
Detailed
Detailed
system
system
architecture
architecture
(CC
(CC
scheme
scheme
)
)
Recognition process
Biometric Application System
Match
Identity
Data
Identity
Data
Extract
Capture
Template
Database
Biometric
character-
istic
Individual
Identity claim
Score
threshold
Pre-
Processing
Feature
Extraction
Raw
data
Comparison
Query
Template
Create
Template
Reference
template
Identity
Data
Identity
Credential
Verification
Identification
Physical signal presentation
Measurement
Identity
Database
Verification mode
Enrollment
Recognition
Comparison
decision
Biometrics
Biometrics
,
,
Authentication
Authentication
7
7
©
©
2008 M
2008 M
ü
ü
ller
ller
Berne University of
Berne University of
Applied Sciences
Applied Sciences
Illustration of
Illustration of
processing
processing
steps
steps
(FP
(FP
recognition
recognition
)
)
Biometrics
Biometrics
,
,
Authentication
Authentication
8
8
©
©
2008 M
2008 M
ü
ü
ller
ller
Berne University of
Berne University of
Applied Sciences
Applied Sciences
Measurement
Measurement
hierarchy
hierarchy
Presentation
Presentation
Attempt
Attempt
Transaction
Transaction
Biometrics
Biometrics
,
,
Authentication
Authentication
9
9
©
©
2008 M
2008 M
ü
ü
ller
ller
Berne University of
Berne University of
Applied Sciences
Applied Sciences
Identification
Identification
(1:n)
(1:n)
Biometrics
Biometrics
,
,
Authentication
Authentication
10
10
©
©
2008 M
2008 M
ü
ü
ller
ller
Berne University of
Berne University of
Applied Sciences
Applied Sciences
Verifikation (1:1)
Verifikation (1:1)
Alleged
Alleged
user
user
Biometrics
Biometrics
,
,
Authentication
Authentication
11
11
©
©
2008 M
2008 M
ü
ü
ller
ller
Berne University of
Berne University of
Applied Sciences
Applied Sciences
Recognition
Recognition
based
based
on
on
raw
raw
data
data
reference
reference
templates
templates
Biometrics
Biometrics
,
,
Authentication
Authentication
12
12
©
©
2008 M
2008 M
ü
ü
ller
ller
Berne University of
Berne University of
Applied Sciences
Applied Sciences
Recognition
Recognition
based
based
on
on
feature
feature
reference
reference
templates
templates
Biometrics
Biometrics
,
,
Authentication
Authentication
13
13
©
©
2008 M
2008 M
ü
ü
ller
ller
Berne University of
Berne University of
Applied Sciences
Applied Sciences
Recognition
Recognition
based
based
on
on
protected
protected
reference
reference
template
template
Biometrics
Biometrics
,
,
Authentication
Authentication
14
14
©
©
2008 M
2008 M
ü
ü
ller
ller
Berne University of
Berne University of
Applied Sciences
Applied Sciences
Recognition
Recognition
based
based
on
on
encapsulated
encapsulated
templates
templates
Biometrics
Biometrics
,
,
Authentication
Authentication
15
15
©
©
2008 M
2008 M
ü
ü
ller
ller
Berne University of
Berne University of
Applied Sciences
Applied Sciences
Quality
Quality
parameters
parameters
(
(
Verification
Verification
)
)
Rate of wrongly accepted non authorized
users in a recognition process
False Accept Rate
FAR
Rate of wrongly refused authorized users in a
recognition process
Rate of the templates from a biometric
characteristic that can be attributed to a
foreign reference template
Rate of the templates from a biometric
characteristic that can not be attributed to its
reference template
Rate of failures for the capture of a template
for a valid recognition process
Rate of population that can not execute an
enrollment process
Failure to Acquire
FTA
False Non Matching
Rate
FNMR
False Match Rate
FMR
False Reject Rate
FRR
Failure to Enrol
FTE
Biometrics
Biometrics
,
,
Authentication
Authentication
16
16
©
©
2008 M
2008 M
ü
ü
ller
ller
Berne University of
Berne University of
Applied Sciences
Applied Sciences
Output of
Output of
comparison
comparison
process
process
Frequency density
s =
s =
Matchingscore
Matchingscore
Biometrics
Biometrics
,
,
Authentication
Authentication
17
17
©
©
2008 M
2008 M
ü
ü
ller
ller
Berne University of
Berne University of
Applied Sciences
Applied Sciences
FNMR, FMR in
FNMR, FMR in
function
function
of
of
decision
decision
threshold
threshold
T
T
s =
s =
Matchingscore
Matchingscore
Density function of score parameters from comparison of templates coming from different biometric characteristics
Density function of score parameters from comparison of templates coming from the same biometric characteristic
1
S
T
Frequency density
FMR
FNMR
ZFR
ZFA
Biometrics
Biometrics
,
,
Authentication
Authentication
18
18
©
©
2008 M
2008 M
ü
ü
ller
ller
Berne University of
Berne University of
Applied Sciences
Applied Sciences
EER
EER
Error rate at thepoint whereFNMR=FMREqualError Rate
EER
Biometrics
Biometrics
,
,
Authentication
Authentication
19
19
©
©
2008 M
2008 M
ü
ü
ller
ller
Berne University of
Berne University of
Applied Sciences
Applied Sciences
Relationship
Relationship
Biometrics
Biometrics
,
,
Authentication
Authentication
20
20
©
©
2008 M
2008 M
ü
ü
ller
ller
Berne University of
Berne University of
Applied Sciences
Applied Sciences
ROC
ROC


Receiver
Receiver
Operating
Operating
Characteristic
Characteristic
Curve
Curve
Theperformanceof a biometricsystemfordifferent decisionthresholdsisrepresentedbythe
ROC (ReceicerOperatingCharacteristic) curve
Thecurveshowsthedependanceof thetwoerrorratesputin functiononeto theother.
(Theoriginal nameof thisplotisDEC (DetectionError Tradeoff) curve, butoftenthetermROC curveisused
althoughthisoriginallynamesa slightlydifferent representationof theerrors.)
Biometrics
Biometrics
,
,
Authentication
Authentication
21
21
©
©
2008 M
2008 M
ü
ü
ller
ller
Berne University of
Berne University of
Applied Sciences
Applied Sciences
Operating
Operating
regions
regions
ROC Kurve
ROC Kurve
Forensics
Forensics
Civil
Civil
applications
applications
High
High
security
security
Biometrics
Biometrics
,
,
Authentication
Authentication
22
22
©
©
2008 M
2008 M
ü
ü
ller
ller
Berne University of
Berne University of
Applied Sciences
Applied Sciences
Example
Example
of ROC
of ROC
curves
curves
for
for
FP
FP
-
-
Systems
Systems
Biometrics
Biometrics
,
,
Authentication
Authentication
23
23
©
©
2008 M
2008 M
ü
ü
ller
ller
Berne University of
Berne University of
Applied Sciences
Applied Sciences
Unique
Unique
application
application
of
of
biometrics
biometrics
PositivRecognition: Verification, Identification
⇒Mapping of a person to its identity
⇒Main security problem: FAR, FMR
⇒Biometrics is an additional factor in an authentication process
(Password, PIN, Token etc)
⇒User are assumed to be cooperative
NegativRecognition: Denial
⇒Separation of a person from an identity, prove of a denied identity
⇒Main security problem : FRR, FNMR
⇒Biometrics is the only factor in a negativrecognition process
⇒User are mostly not cooperative
Biometrics
Biometrics
,
,
Authentication
Authentication
24
24
©
©
2008 M
2008 M
ü
ü
ller
ller
Berne University of
Berne University of
Applied Sciences
Applied Sciences
Threads
Threads
to a
to a
biometric
biometric
system
system
Biometrics
Biometrics
,
,
Authentication
Authentication
25
25
©
©
2008 M
2008 M
ü
ü
ller
ller
Berne University of
Berne University of
Applied Sciences
Applied Sciences
Problems of
Problems of
biometrics
biometrics
Administration: Enrolment Prozess
Jurisdiction: Non acceptance in court, data
protection
Security:Protection of infrastructure,
Tampering
TCO:Measurement system, degradation
Scaling:(n-n) Matching, template overlap
Profiling:Traceable biometric data
Data corruption:Biometric characteristics are
unchangeable
Proprietary Systems:Standards (JTC1/SC37/BioAPI/CC)
Counterfeit technology:Unclear risks, mathematically
difficult to define
Biometrics
Biometrics
,
,
Authentication
Authentication
26
26
©
©
2008 M
2008 M
ü
ü
ller
ller
Berne University of
Berne University of
Applied Sciences
Applied Sciences
Reduced
Reduced
threads
threads
and
and
risks
risks
:
:
Encapsulated
Encapsulated
biometrics
biometrics
(
(
Encapsulated
Encapsulated
Biometrics
Biometrics
)
)
Biometrics
Biometrics
,
,
Authentication
Authentication
27
27
©
©
2008 M
2008 M
ü
ü
ller
ller
Berne University of
Berne University of
Applied Sciences
Applied Sciences
Biometric
Biometric
System
System
Definition:
„Biometrics is a pattern recognition system that recognizes persons
by some characteristic physiologic or behaviorist features.“
Attribute: mandatory
Universal:All persons have the feature
Distinctive:Each person has a distinct feature
Long lived:Features are invariant over the time
Measurable:Feature can be measured
Attribute: optional
Quality:Feature is simple to measure, separates maximal
Acceptance:Persons are willing to accept the measurement
Fraud:It is difficult to fool the measurement system
Biometrics
Biometrics
,
,
Authentication
Authentication
28
28
©
©
2008 M
2008 M
ü
ü
ller
ller
Berne University of
Berne University of
Applied Sciences
Applied Sciences
Overview
Overview
on
on
common
common
biometric
biometric
features
features
Physiological features
Finger print
Iris
Retina
Veins
Palm
Face
Ear form
Finger geometries
DNA, Protein
Odor
Temperature image (hand,
face)
Lip print
Teeth bit
…….
Behaviorist features
Voice
Hand writing
Hand movement dynamics
Gait
Keyboard pressure dynamics
Grip
………
Biometrics
Biometrics
,
,
Authentication
Authentication
29
29
©
©
2008 M
2008 M
ü
ü
ller
ller
Berne University of
Berne University of
Applied Sciences
Applied Sciences
Most
Most
important
important
biometric
biometric
characteristics
characteristics
Biometrics
Biometrics
,
,
Authentication
Authentication
30
30
©
©
2008 M
2008 M
ü
ü
ller
ller
Berne University of
Berne University of
Applied Sciences
Applied Sciences
Wichtigste Techniken
Wichtigste Techniken
Biometrics
Biometrics
,
,
Authentication
Authentication
31
31
©
©
2008 M
2008 M
ü
ü
ller
ller
Berne University of
Berne University of
Applied Sciences
Applied Sciences
Qualitativ
Qualitativ
valuation
valuation
Biometrics
Biometrics
,
,
Authentication
Authentication
32
32
©
©
2008 M
2008 M
ü
ü
ller
ller
Berne University of
Berne University of
Applied Sciences
Applied Sciences
Quantitativ
Quantitativ
valuation
valuation
Telephone
service
Low1/30VoicecharacteristicsVoiceprinting
Low-security
facilities
Low1/100
Shapeof letters, writingorder, pen
pressure
Signature
Low-security
facilities
Low1/100
Outline, shapeand distributionof eyes
and nose
Facial
Recognition
Low-security
facilities
Low1/700Size, lengthand thicknessof handsHand Shape
UniversalMedium1/1,000FingerprintsFingerprinting
High-security
facilities
High1/1,200,000Iris pattern
Iris
Recognition
Applications
Secur-
ity
Misidentification
rate
CodedPatternMethod
Source: AIM Japan, Automatic IdentificationSeminar, Sept.14, 2001
Biometrics
Biometrics
,
,
Authentication
Authentication
33
33
©
©
2008 M
2008 M
ü
ü
ller
ller
Berne University of
Berne University of
Applied Sciences
Applied Sciences
Economical
Economical
development
development
Biometrics
Biometrics
,
,
Authentication
Authentication
34
34
©
©
2008 M
2008 M
ü
ü
ller
ller
Berne University of
Berne University of
Applied Sciences
Applied Sciences
Market Share
Market Share
by
by
Technology
Technology