GA10 People and Security Lecture 3: Biometrics Applying risk - - PDF document

GA10 People and Security Lecture 3: Biometrics

Applying risk analysis

Goals of this lecture

• Brief introduction to biometrics
• Context: ICAO,US Visit and the ID cards

programme

• Issues with current equipment

– Performance – Usability – User Acceptance

• How secure are biometrics?

Applying risk analysis

Basics on biometrics

• Enrolment and

subsequent

– verification (through ID + biometric), or – identification (biometric only)

• Full images or templates

– Passports requires images, templates are more efficient

• Size of database affects

performance

Applying risk analysis

Physical biometrics

• Fingerprint
• Finger / Palm Vein
• Hand geometry
• Face recognition
• Iris
• Retina
• Earshape

Applying risk analysis

Behavioural biometrics

• Voice print
• Dynamic Signature Recognition (DSR)
• Typing pattern
• Gait recognition
• Heart rate analysis

Applying risk analysis

Fingerprint recognition

• Applications

– Authentication/Access control

• Doors
• PCs/laptops
• US Visit programme

(http://www.dhs.gov/dhspublic/interapp/content_multi_image /content_multi_image_0006.xml)

Applying risk analysis

Hand geometry

• Applications

– Authentication (e.g. INSPASS program)

• Usability

– Easier to position hand than fingers – Less susceptible to small injuries – Hygiene concerns

Applying risk analysis

Voice recognition applications

• Applications

– Speaker recognition – Telephony-based interactions (home banking and insurance) – Lie detector

• Usability issues

– Speaker training – Voice changes – colds etc. – Background noise

Applying risk analysis

Dynamic Signature Recognition

• Applications

– Electronic documents with signature: contracts, mortgage agreements – Anything that needs signing

• Usability issues

– Natural interaction that most users understand, but difficult on handhelds – Declaration of will

Applying risk analysis

Biometrics Authentication

Applying risk analysis

Enrolment

• Crucial for security and subsequent

performance

– In some context, identity of enrolee needs to be checked – Biometrics enrolled need to be

• genuine (see attacks)
• good enough quality to work
• Enrolment procedure needs to be formalised

– Staff need to be trained – Staff need to be trustworthy or closely checked

• Time taken to carry out enrolment often

under-estimated

Applying risk analysis

Failure to Enrol (FTE) & Failure to Acquire (FTA)

• FTEs and FTAs threaten Universal Access
• Reasons for FTE/FTA

– Biometric not present – Biometric not sufficiently prominent or stable

• Finger

– wearing down of fingerprints, callouses (manual work, chemicals, sports, age), deformation, arthritis

• Iris

– missing iris, very dark eyes, glasses or contacts (reflection or frame), drooping eyelids

• Face

– veils, eyepatches, headcoverings, severe disfigurement, inability to keep still

Applying risk analysis

Context

• International developments

– ICAO agreement – US Visit

• UK ID legislation

– Stand-alone ID card for everyone over 16 – 3 biometrics (face, 10 finger, 2 iris) on card, and in National Identity Register – Access by govt departments, federated ID – Access by commercial organisations

Applying risk analysis

Example FTE rates from UKPS enrolment trial

3.91% 39% 2.73% Disabled 0.69% 12.30% 0.15% Quota Finger Iris Face

Applying risk analysis

False Acceptance Rate (FAR) & False Rejection Rates (FRR)

• FAR

– accepting user who is not registered – mistaking one registered user for another – ICAO: FAR of .01% is regarded as acceptable

• FRR
• – rejecting registered user
• High FRRs reduce usability, high FARs reduce

security – customer-based applications tend to raise FAR

Applying risk analysis

Performance

• User performance depends on

– frequency of use:

• Frequent users complete faster and with fewer errors,

infrequent users need step-by-step guidance and detailed feedback

– Degree of cooperation – Total usage time (not just for matching)

Applying risk analysis

Applying risk analysis

Applying risk analysis

"W e w ere aim ing for it to scan 1 2 pupils a m inute, but it w as only m anaging 5 so has been tem porarily suspended as w e do not w ant pupils' m eals getting cold w hile they w ait in the queue."

Careful balancing of business process requirements and security requirements needed

Applying risk analysis

Total Usage Process

• Time quoted by suppliers often only refer to capture
• f live image & matching

– Walk up to machine – Put down bags, remove hats, etc. – Find token (if used) – Put on token (if used) – Read token – Wait for live image to be captured & matched – Walk away & free machine for next user – Plus average number of rejections & re-tries

Average usage time in BioPII 12-20 seconds, longer with infrequent users

Applying risk analysis

FRR rates from UKPS enrolment trial

16.35%

1 min 20 sec

8.22%

1 min 18 sec

51.57%

1 min 3 sec

Disabled 11.70%

1 min 13 sec

1.75%

58 sec

30.82%

39 sec

Quota Finger Iris Face

Applying risk analysis

Example: Disney Orlando

• Goal: revenue protection
• Technology: hand geometry
• Users: season ticket holders (4000)
• Performance:

– High FAR threshold (5% +) – Soft response to rejections – 9-10 secs, ops people grumble: 5 secs needed

Applying risk analysis

Example: Smartgate Sydney Airport

• Problem: speedy & secure immigration
• Technology: Face recognition system
• Users: Quantas air crew (2000)
• Performance:

– FAR “less than 1%” – FRR 2% – “could be faster” (average 12 secs)

• Several re-designs necessary, including updating of image

templates

Applying risk analysis

Usability Issues: Finger

• Which finger?
• How to position

– Where on sensor? – Which part of finger? – Straight or sideways?

• Problems: arthritis, long fingernails, handcreme,

circulation problems

Applying risk analysis

Which finger?

Applying risk analysis

Finger position?

Applying risk analysis

Usability Issues: Iris

• What is it – iris or face?
• One or both eyes?
• One eye: how to focus?
• Distance adjustment
• Positioning

– “rocking” or “swaying”

• Glasses and contact lenses

– about half of population wear them – Target area difficult to see when glasses are removed

Applying risk analysis

Focussing

Applying risk analysis

Height adjustment

• Often not sufficient for very short (under 1.55 m)
• r very tall (over 2.10) people, or wheelchair users
• Need to use hand to adjust

– If card needs to be held, other things users carry or hold need to be put down

Applying risk analysis

Height adjustment

Applying risk analysis

… but users may not realise this

… or be reluctant to touch equipment,

• r think it takes too long

Applying risk analysis

Usability Issues: Face

• What is it?
• Where do I stand?
• Where do I look/what am I looking at?
• Standing straight, keeping still
• “Neutral expression”
• Hats, changes in (facial) hair, makeup

Applying risk analysis

Distance

Applying risk analysis

“Neutral expression”

Applying risk analysis

UK Passport Service Trial

• Best performing: iris with “normal” users – FRR

4%

• Worst performing: face recognition with disabled

users - FRR 30%

• Verification time: 40-80 secs
• With a database of 10.000 people

Applying risk analysis

User Acceptance

• Acceptance requires

– perceived need for security – trust in operator – convenience, or at least usability

Applying risk analysis

User Acceptance Issues –Finger

• Hygiene, Hygiene,

Hygiene

• Association with

forensics/criminals

• Finger chopped off

Applying risk analysis

Applying risk analysis

Applying risk analysis

Liveness detection

• Detects movement, pulse, blood flow
• Fitted to several systems, but tends to increase

FRR

• Users: fine, but do the criminals know about it?

Applying risk analysis

User Acceptance Issues - Iris

• Iris

– Risk to health (e.g. damage to eyes, triggering epilepsy) – Covert medical diagnosis

• Illnesses (iridiology)
• Pregnancy
• Drugs
• “Minority Report” attacks

Applying risk analysis

User Acceptance Issues - Face

• Covert identification
• Surveillance/tracking

– Direct marketing

Applying risk analysis

User Acceptance – General Issues

• Data protection – threat to privacy
• Abuse by employer, commercial organisations,

state, or malicious individuals

– Increasing capability of technology – e.g. iris recognition at a distance – Integration with other technologies – e.g. RFID

• Doubts about reliability

– Sophisticated attackers – Can governement really keep systems secure? – Cheap systems and successful attacks erode confidence

Applying risk analysis

Applying risk analysis

Applying risk analysis

• Simple

– Activate latent prints: breathing, bag with warm water

• Sophisticated

– Lift print with tape or photograph

• Gelatine print

(gummy bear attack) – lasts 1x

• Silicone print

Attacks - Finger

Applying risk analysis

Attacks - Iris

• Simple

– Picture of eye stuck on glasses

• Sophisticated

– Coloured contact lense

• Get someone else to

enrol and give you lenses & passport

• Print iris image onto

coloured contact lense

Applying risk analysis

Attacks - Face

• Simple

– Photo or video of person – Glasses

• Sophisticated

– Mask (Mission Impossible attack) http://www.heise.de/ct/english/02/11/114/bild7.jpg

Applying risk analysis

Applying risk analysis

Maintaing performance in continuing use

• Maintenance required to maintain performance,

e.g. cleaning of sensors

• Keeping systems secure from attacks (including

vandalism and insider attacks)

• Secure and efficient processes for dealing with

FRRs and FTE/FTA users

• Secure & efficient processes for new enrolments,

and secure re-enrolment in cases with high FRR

• r significant changes

Applying risk analysis

Summary

• Biometrics have potential to reduce

users’ workload and improve business process, IF

• 1. systems are reliable, robust, easy to install,

maintain and use

• 2. Security of overall system can be assured
• 3. performance meets required level and can be

sustained in everyday use

• 4. Systems are accessible and acceptable to

users.

Applying risk analysis

Challenges

• Large public applications face additional issues
• Universal access

– System with best FTE performance (face) has worst FRR – Systems with better FRR have unacceptable FTE (for some applications) – Accommodating all users to provide universal access