Research in Liveness Detection - Martin Drahanský

Nadpis 1Nadpis 2Nadpis 3

Jméno PříjmeníVysoké učení technické v Brně, Fakulta informačních technologií v Brně

Božetěchova 2, 612 66 [email protected]

99.99.2008

Research in Liveness Detection

Martin DrahanskýBrno University of Technology, Faculty of Information Technology

Božetěchova 2, 612 66 Brno, Czech Republichttp://www.fit.vutbr.cz/~drahan

[email protected]

April 11, 2015© Presentation with copyright (Security Session 2015)

Technologies of fingerprint sensors

Research in Liveness Detection

• Optical technology

• Capacitive technology

• Ultrasound technology

• E-field technology

• Electro-optical technology

• Pressure sensitive technology

• Thermal technology

2/29

Security of biometric systems

SensorFeature

ExtractionMatcher

Enrollment Center

StoredTemplates

Yes

/No

1 2 3 4

5

6 7

8

9

10

11

Research in Liveness Detection 3/29

Finger fakes

Suprema SFM 3020(fake – rubber stamp)

Artificial fingerprints (SFinGe)

Artificial fingers

Fake fingerprints of different materialsResearch in Liveness Detection 4/29

How to produce a fake finger(print)?

1 2 3 4

5 6 7 8

9 10 11 12

P l a y v i d e o

Research in Liveness Detection 5/29

Perspiration

• Clarkson/Virginia universities

• 600 sweat glands in 1 inch2

• Duration approx. 5 seconds

• High intra-class variability

Time

Research in Liveness Detection 6/29

Spectroscopic characteristics

Illumination

Imaging

SpectroscopicSpoof Detector

Classical FTIR Principle

• Lumidigm Ltd.(Albuquerque)• Clones: TST Biometrics GmbH, Sagem Morpho etc.

Research in Liveness Detection 7/29

Ultrasonic technology

• Ultra-Scan / Optel

Sound wavepulse transmission

Sound wave echos,captured to produce images:

Echo #1

Echo #2

Echo #3

Desired image depthis selected by range gate

Ridgestructure

Air gap or contamination

Platen

Research in Liveness Detection 8/29

Temperature + Hot & cold stimulus

• Temperature

• Hot and cold stimulus

Research in Liveness Detection 9/29

Pressure stimulus

• National utility model ÚPV-19364

• Ridge elasticity• 20%

• Change of color (RGB)• G ~ 42

Research in Liveness Detection 10/29

Pulsation of pixels

• Using RGB, each channel

shown tiny pulsation

• Could be used for mobile

devices

• Only transparent finger

fakes could be used, all

other do not work

• Our patent

ÚPV-304801

Research in Liveness Detection 11/29

Electrical properties

• Bio-impedance

• Resistance / Conductivity

ohm

0 200 400 600 800 1000

kΩ

L_Thumb

L_Index

L_Middle

L_Ring

L_Little

R_Thumb

R_Index

R_Middle

R_Ring

R_Little

Fin

ger

Measurement with DC low voltage

Research in Liveness Detection 12/29

Pulse I.

• Heart activity

Research in Liveness Detection 13/29

Pulse II.

• International patent WO/2007/036370

• National patent pending; new approach in preparation

Common optical fingerprint scanner

High resolution camera with

macro-objective

Laser distance measurement

module

Option 1

Option 2

Research in Liveness Detection 14/29

Pulse III.

• Macroobjective I.

Glass

Finger (front view)

Mac

ro-o

bjec

tive

CCD camera

CC

D c

amer

aSemi-permeablemirror

Research in Liveness Detection 15/29

Pulse IV.

• Macroobjective II.

-4

-2

0

2

4

6

8

10

12

14

1 7 13 19 25 31 37 43 49 55 61 67 73 79 85 91 97 103 109 115 121 127 133 139 145 151 157 163 169 175 181 187 193 199 205 211 217

Image number

Dis

tan

ce c

han

ge

[pix

.] Real heart beats: 76 / min

23 peaks in 15 sec correspond to 92 heart beats in 1 min

Research in Liveness Detection 16/29

Pulse V.

• Laser I.

CCD camera

Finger(front view)

Gla

ss

Lase

rm

odul

e

Glass

Research in Liveness Detection 17/29

Pulse VI.

1.9 subdiv

6 su

bdiv

25 periodical runs in 20 seconds

Research in Liveness Detection 18/29

75 heart beats in 1 min. (this method)77 heart beats in 1 min. (tonometer)

Troubles during scanning process

Ber

gdat

a F

CA

T-1

00

Earth dust

Ver

idic

om 5

th S

ense

Metallic dust Fine sand Oiled finger

Ber

gdat

a F

CA

T-1

00About -10°C

Ver

idic

om 5

th S

ense

About +50°C

• Change of temperature and/or humidity

• Dry or wet fingers

• Dirt on sensor area

• Dirt on fingertip surface

• Blurred images

• Change of pressure

• Latent fingerprints etc.

Research in Liveness Detection 19/29

Experiments with real fingers and fakes

Research in Liveness Detection 20/29

Skin diseases I.

• Histopathological changes

• Changes of skin color

• Histopathological and skin color changes

Staphylococcal abscess

Eczema

Pyogenic granuloma Scleroderma

HFM En

do

ca

rditis

Tinea Hy

pe

rke

rato

sis

Eczematous inflammation Irritant dermatitis Pompholyx

Psoriasis

Research in Liveness Detection 21/29

Skin diseases II.

Project COST

Research in Liveness Detection 22/29

Finger veins – Principle

Veins

Arteries

Ab

sorp

tio

n

Research in Liveness Detection 23/29

Finger veins – Our Pre-α-Prototype

IR LEDs

Black-Box with a small door

IR-Filter

Camera Sony (VGA, Low-Lux)

Screwdriver

Missing: Polarization-Filter

Research in Liveness Detection 24/29

Finger veins – Acquired image

Research in Liveness Detection 25/29

Finger veins – Image processing

a) b) c) d) e) f)

Research in Liveness Detection 26/29

Eye iris and eye retina – Basic principles

• One device for acquirement of eye iris and eye retina → possible ways of use are for biometrics and ophthalmology

Research in Liveness Detection 27/29

Anti-spoofing in 3D face recognition

• We use printed and cut 3D face fakes

Research in Liveness Detection 28/29

Anti-spoofing in thermo-face recognition

FLIR ThermaCAM E300

Optris PI200

http://astro.swarthmore.edu/ir/ir_basics.html

Research in Liveness Detection 29/29

Skalský dvůr 2009 30

Thank you for your attention.