Copyright © 2001 Visionics Corporation FaceIt®: Enabling Technology for Airports and Border Control.

Copyright © 2001 Visionics Corporation

FaceIt®:

Enabling Technology for

Airports and Border Control

Presentation Overview

– Corporate Snapshot

– Choosing a Biometric

– Benefits of Facial Biometrics

– Border Control Applications

– Sample Installations

– Biometric Performance

– Biometric Fusion

VISIONICS & IDENTIX AGREE TO A MERGER OF EQUALS

• Merger Announced 2/22/02• Now, World’s Largest Multi-Biometric

Provider of Facial Recognition and Fingerprint technologies

• Will Enable the Offering of Complete, Integrated Biometric Solutions

RECENT ANNOUNCEMENT

Visionics: Quick Facts

UK

CaliforniaNew Jersey

Minnesota

Nasdaq Symbol VSNX

~ 200 employees

FY01 (E): $30M revenue

Worldwide biometrics industry leader

Founding member of the International Biometric Industry Association (IBIA)

Lines of BusinessLines of Business

Lines of Business

4. IBIS4. IBIS

Mobile Mobile identificatiidentification systems on systems

for law for law enforcemeenforceme

ntnt

3. Live 3. Live ScanScan

Electronic Electronic capture of capture of forensic forensic quality quality

fingerprinfingerprint imagest images

2. BNA2. BNA

Hardware Hardware appliances appliances & modules & modules

that that encapsulatencapsulat

e e biometric biometric capabilitiecapabilitie

ss

1.FaceIt1.FaceIt

Face Face recognitiorecognitio

n n software software technologtechnolog

yy

• Developer of FaceIt® Developer of FaceIt® engines.engines.

• Focused on building Focused on building cutting edge cutting edge technology modules technology modules that are used by that are used by System integrators, System integrators, product developers product developers and OEMs.and OEMs.

• Founded by scientists Founded by scientists who discovered the who discovered the basic algorithms. basic algorithms. Formerly at Princeton Formerly at Princeton and Rockefeller and Rockefeller Universities.Universities.

Revenue leaders in facial recognition

Most deployments

Most partners

Worldwide brand name recognition

Leaders in Face Recognition

Biometrics

Face recognition is a special example of a

Biometrics

Biometric Technology

BiometricsBiometrics

Biometric Technology

Computerized methods for automatically recognizing a person based on a physical trait or attribute.

The Biometrics Family

I. High Intrinsic Accuracy

II. Limited Intrinsic Accuracy

III. Commercially Nonviable

. . .

The Biometrics Family

The Facial Promise

A photograph-- whether static or a live image---contains the information necessary to distinguish one person among millions of others—FaceIt® Faceprint.

The Facial Promise

A photograph-- whether static or a live image---contains the information necessary to distinguish one person among millions of others—FaceIt® Faceprint.

Security with a Human Face™

A powerful software tool!

Why Face?

I. FamiliarIDs have used photos for years.

II. CompatibleWorks with existing ID dbases..

III. SimpleEasy to use. User does nothing!

IV. Human-backupHumans read face not finger or iris

Why Face?

Different Reasons in Different Apps:

Why Face?

V. Audit trailHuman-Readable

VI. ID at a Distance

Surveillance

How Does Face Recognition Work?

Making faces

Local Feature Analysis:Local Feature Analysis: Approach used by Visionics.

Local Analysis.

Principal Component Analysis:Principal Component Analysis: Approach used by competition (Eigenfaces).

Global Analysis.

Most fundamental element in face recognition is the representation used to build the digital code (i.e. faceprint).

Making faces

Another approach that was considered is Neural Networks but was rapidly discredited within the scientific community as very inaccurate.

Local Feature Analysis

Nodal Points: Faces, like finger prints, possess Nodal Points that can be automatically detected.

Identity: Geometry of constellation of Nodal Points is unique to a given face and is over

redundant.

Local Feature Analysis

Right…

FrontotemporaleGlabella Sellion

Pupila Zygofrontale Zygion Infraorbitale Ectocanthus Infrazygion Inframalar Tragion Gonion Pronasale MIM Promenton Menton …

Left…

FrontotemporaleGlabella Sellion

Pupila Zygofrontale Zygion Infraorbitale Ectocanthus Infrazygion Inframalar Tragion Gonion Pronasale MIM Promenton Menton …

Local Feature Analysis

Properties of LFA

Color Blind: Does not use color

information. Increased robustness with respect to

changes in skin tone.

Racially Blind: Performance is independent of race. Facial structure is universal across races.

Versatile Input: Can work from any image source

including near-infrared or night vision.

Robust with Aging: The faceprint maintains a significant portion of its

structure as the face ages.

Properties of LFA

Visionics technology can provide three key applications within the airport and border crossing environment…

Livescan biometric Livescan biometric systemssystems

Capture NIST records

Electronic submission to IAFIS /NCIC

Channeling AgenciesChanneling Agencies

Office of Personnel Management

AAAE

Indian Gaming Commission

Nuclear Regulatory Commission

40 40 millionmillion

RecordsRecords

1. Background Checking: Already in Place

System Example: Airport System Example: Airport Applicant Systems: Feb Applicant Systems: Feb

0202

Visionics and Identix are currently installed in over 100 US Airports.Visionics and Identix are currently installed in over 100 US Airports.

Visionics provides over 600 live scan systems at INS Applicant Visionics provides over 600 live scan systems at INS Applicant Support Centers across the USSupport Centers across the US

Pre 9/11

Post 9/11

Granting of access for a variety of privileges:

Expedited border clearance

Expedited airplane boarding - SPT

Secure access: facilities

Trusted Identity data need NOT reside in a national database

Entity administering privilege—commercial & government—can maintain, separate archives (privacy concerns)

Dual or multi-use tokens (i.e. smart cards)

2. Privileged Access: One-to-One Matching

•System Integrator EDS was awarded a contract by the Israeli Government for a Biometric facilitated Border Control application.

40,000 daily workers to enter and exit the Gaza Strip via 42 routing passages in “terminal-like” environment.

The project will introduce the use of Fusion Module for the use of two biometrics.

Example: Basel – Border Control Application

Privileged Access: One-to Many

A Task that Cannot be done

Manually.

1. At enrollment, automatically check uniqueness of Identity.

2. Eliminates duplicates & Aliases.

Additional System Examples:

Colorado State, DMV

West Virginia, DMV

Mexican National Voter Election System

Terror and crime are not faceless

Improved intelligence is yielding the identities of perpetrators

Three Primary Applications:

ID-document surveillance

Checkpoint surveillance

On-Demand surveillance

3. Watch-List Detection

ID-Document Surveillance Uses standard

existing travel documents

Creates manifest: entry-exit

System Example:

Dominican Republic - First large scale implementation

120 points of entry

Facial images from ID documents scanned upon entry & checked against centralized terrorist, drug traffickers & criminals watchlists

AlertAlert

Typical BNA Configuration: FaceIt® ARGUS

Checkpoint Surveillance

FaceIt® Bio TemplateFaceIt® Bio Template

FaceIt® Bio CaptureFaceIt® Bio Capture

FaceIt® Bio CaptureFaceIt® Bio Capture

FaceIt® Bio CaptureFaceIt® Bio Capture

FaceIt® Bio TemplateFaceIt® Bio Template

FaceIt® Bio TemplateFaceIt® Bio Template

FaceIt® Bio LogicFaceIt® Bio Logic

FaceIt® Bio SearchFaceIt® Bio Search

Capable ofWireless alertsclosest guards

User WorkstationUser Workstation Control,Control,

review, administrationreview, administration

Current Airport Installations:

Fresno Yosemite Airport

Keflavik International Airport, Iceland

Dallas Ft. Worth

Palm Beach Airport

Boston Logan (Testing Complete)

Several unnamed airports & border crossing (under installation)

Also installed in:

Newham, London, UK

Birmingham, UK

Ybor City, Tampa, Florida

Sample Surveillance Installations

On-Demand Surveillance: IBIS

System Examples:System Examples:

West Valley, UtahWest Valley, Utah

Hennepin County, Hennepin County, MNMN

Redlands, CARedlands, CA

Ontario, CAOntario, CA

The accuracy of all biometrics can be assessed via:

False Accept Rate (FAR),

False Reject Rate (FRR),

Failure to Acquire (FTA).

FAR = Risk

FTA = Liability

Golden Question: How Accurate is it?

Technology has limitations, must build systems that take them into accounts

Setting realistic expectations: (will never do FRR=0, FAR=0, FTA=0, cost=0)

Exception Handling:

Secondary screening with human backup

Fusion approach

Managing Customer/User Expectations

Fusion versus Layering

Score Fusion

(statistical properties of

multiple biometrics)

Layering:AND OR

BiometricsModule 1

Decision One

BiometricsModule 2

Decision Two

BiometricsModule 1

BiometricsModule 2

Score One

Score Two

Decision

Final Decision

FARFAR 0.1%0.1% 0.5%0.5% 1%1%

FRR: B1FRR: B1 3.6%3.6% 2.0%2.0% 0.6%0.6%

FRR: B2FRR: B2 0.23%0.23% 0.11%0.11% 0.08%0.08%

FRR: “OR”FRR: “OR” 2.5%2.5% 1%1% 0.63%0.63%

FRR: “AND”FRR: “AND” 0.9%0.9% 0.56%0.56% 0.048%0.048%

FRR: FUSIONFRR: FUSION 0.004%0.004% 0.001%0.001% 0.00070.0007%%

Fusion: Improvements in Accuracy

Recently Passed Legislation

• Aviation and Transportation Security Act of 2001

Established the Transportation Security Administration (TSA). Mandates background checks on all airport employees. Also mandates upgraded access control systems for secure areas at airports.

• USA Patriot Act of 2001

Expands government surveillance powers. Requires technology standard for tracking entry and exit of aliens. Suggests the use of biometric technology to conduct background checks prior to issuance of visas.

Currently Pending Legislation

• Drivers License Modernization Act of 2002

Would require a biometric to be embedded on an individual's driver's license. Requires that the Department of Transportation develop a technology standard in conjunction with other governing bodies

• Enhanced Border Security and Visa Entry Reform Act Mandates the creation of a watch-list database, made

available to all agencies involved in determining the admissibility and the identification of aliens. Mandates the issuance of machine-readable, tamper-resistant travel documents with biometric identifiers and the installation of equipment and software capable of reading the new documents at all U.S. ports of entry by October 26, 2003.

Copyright, 2001 © Visionics Corporation.

Thank You For Your

Attention.For further information,

please contact:Vaughn J. Barber

Business Development(201) 332 – 9213 x202

[email protected]