An Introduction to Biometric Identity Verification Gérard CHOLLET [email protected]@ GET-ENST/CNRS-LTCI 46 rue Barrault 75634 PARIS cedex 13 chollet.

An Introduction to Biometric Identity Verification

Gérard [email protected]

GET-ENST/CNRS-LTCI46 rue Barrault

75634 PARIS cedex 13http://www.tsi.enst.fr/~chollet

Outline

What is Biometry ? Why is it important ? Biometric modalities, Physical and behavioral

characteristics Pattern recognition and Decision theory Multimodal Identity Verification Databases, Evaluation, Standardization Applications Introduction to further presentations Perspectives

What is BIOMETRICS ?

This term has several meanings : statistical and mathematical methods applicable

to data analysis problems in the biological sciences

Biometrics are automated methods of recognizing a person based on a physiological or behavioral characteristic. 

The second meaning is of concern here. It is a hot topic for security and prevention of

identity theft

Why is it important to recognizethe identity of a person ?

Identification and/or Verification Protection of individual property (habitation,

bank account, personal data, messages, mobile phone, PDA,...)

Limited access (secured areas, data bases) Locate a particular person in an audio-visual

document (information retrieval) Who is speaking in a meeting ? Is a suspect the criminal ? (forensic

applications)

How to verify the identity of a person ?

Control a specific knowledge (password, PIN,...) or the possession of a document (passport, ID card) or a physical

qui risque d’être oublié par son propriétaire ou usurpé par un imposteur

contrôler une possession (passeport, clé, badge,...) qui risque d’être volé

mesurer les caractéristiques physiques (visage, empreintes digitales, iris,...) ou comportementales (parole, signature,...) de l’individu

une combinaison de ces moyens rend l’imposture difficile mais complique l’accès

Modalities for Identity Verification

A device (key, smart card,…) or a document (passport, ID card) you own

A code you remember (password, …) Could be lost or stolen

Physiological characteristics: Face, iris, finger print, hand shape,…

Need special equipments Behavioral characteristics:

Speech, signature, keystroke, gait,… Speech is the prefered modality over the telephone(but a ‘voice print’ is much more variable than a finger

print)

Modalities for identity verification

Bla-bla

SECUREDSPACE

PIN PIN 1111111111111111

11

Physical Biometric Modalities

Face (visible light, infra-red, thermogram, 3D, …)

Finger print Retinal scan, Iris Hand geometry, Veins, Palm print Ear shape, Genetic code ...

Behavioral Biometric Modalities

Speech (text dependent, text independent, …) Hand writing, signature Gesture, Gait Keystroke pattern on a keyboard …

Desired properties of a Biometric modality

Easy to measure (for real time verification) Efficient (precision, speed, cost) unicity (2 persons should not have identical

characteristics) sustainable (NO temporal drift) User acceptance impossible to duplicate (robustness to forgery)

Empreintes digitales

Empreintes digitales

Minuties

Le visage

Caméra infra-rouge

(a) (b) (c)

(d) (e) (f)

Best-fit ellipse image

Rotation

Normalized imageErosion and sharpening

Simplified image

Gradient image

Adaptive Hough transform and

template matchingSnake energy:

exttotal EEE int

Normalisation du contraste

Initial ImagesInitial Images

After After NormalizationNormalization

Rétine

Localisation de l’iris

Iris

Comparaison des caractéristiques de l’iris

Signatures

La démarche

Speaker Verification

Typology of approaches (EAGLES Handbook) Text dependent

Public password Private password Customized password Text prompted

Text independent Incremental enrolment Evaluation

Architecture d’un système de reconnaissance biométrique

Registration of a new client

Acquisition of biometric patterns to be used as reference.

For a number of modalities (signature, vocal password,...), several repetitions are desired.

A reference model may be infered from the reference patterns.

This model could be adapted to follow temporal drifts.

Recognition of a person

Is he really the person he claims to be ? Identity verification

Who am I ? Identification (the closest person in a closed set) Followed by verification to reject unknown

individuals Deliberate imposture is a major problem in

identity verification

Two types of errors : False rejection (a client is rejected) False acceptation (an impostor is accepted)

Decision theory : given an observation O and a claimed identity H0 hypothesis : it comes from an impostor H1 hypothesis : it comes from our client

H1 is chosen if and only if P(H1|O) > P(H0|O) which could be rewritten (using Bayes law) as

Decision theory for identity verification

)1()(

)(

)1(

HPHoP

HoOP

HOP

)1()(

)(

)1(

HPHoP

HoOP

HOP

Decision for Identity Verification

Likelihood ratio

Distribution des scores

Receiver Operating Characteristic (ROC) curve

Detection Error Tradeoff (DET) Curve

Speaker Verification

Typology of approaches (EAGLES Handbook) Text dependent

Public password Private password Customized password Text prompted

Text independent Incremental enrolment Evaluation

Inter-speaker Variability

We wereaway

ayear ago.

Intra-speaker Variability

We

were

away

a

year

ago.

Dynamic Time Warping (DTW)

HMM structure depends on the application

Signal detection theory

Speaker Verification (text independent)

The ELISA consortium ENST, LIA, IRISA, ... http://www.lia.univ-avignon.fr/equipes/RAL/elisa/

index_en.html

NIST evaluations http://www.nist.gov/speech/tests/spk/

index.htm

Gaussian Mixture Model

Parametric representation of the probability distribution of observations:

Gaussian Mixture Models

8 Gaussians per mixture

National Institute of Standards & Technology (NIST)

Speaker Verification Evaluations

• Annual evaluation since 1995• Common paradigm for comparing technologies

GMM speaker modeling

Front-endGMM

MODELING

WORLDGMM

MODEL

Front-end GMM model adaptation

TARGETGMM

MODEL

Baseline GMM method

HYPOTH.TARGET

GMM MOD.

Front-end

WORLDGMM

MODEL

Test Speech

xPxPLog ]

)/()/([

LLR SCORE

)/( xP

)/( xP

=

Support Vector Machines and Speaker Verification

Hybrid GMM-SVM system is proposed

SVM scoring model trained on development data to classify true-target speakers access and impostors access,using new feature representation based on GMMs

Modeling

Scoring

GMM

SVM

SVM principles

X (X)

Inpu

t sp

ace

Feat

ure

spac

e Separating hyperplans H , with the optimal hyperplan Ho

Ho

H

Class(X)

Results

Multimodal Identity Verification

M2VTS (face and speech) front view and profile pseudo-3D with coherent light

BIOMET:(face, speech, fingerprint, signature, hand shape) data collection reuse of the M2VTS and DAVID data bases experiments on the fusion of modalities

BIOMET

An extension of the M2VTS and DAVID projects to include such modalities as signature, finger print, hand shape.

Initial support (two years) is provided by GET (Groupement des Ecoles de Télécommunications)

Emphasis will be on fusion of scores obtained from two or more modalities.

Presentations to come

Perspectives

A lot of interest from governments, telecom and financial operators,…

Fusion of modalities.

A number of R&D projects within the EU.

Smart cards to support biometric references and to perform identity verification.