BIOMETRIC IDENTIFICATION OF PERSONS · 2017. 4. 9. · BIOMETRIC IDENTIFICATION OF PERSONS Milan Adámek, Petr Neumann, Dora Lapková, Martin Pospíšilík and Miroslav Matýsek Tomas
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
BIOMETRIC IDENTIFICATION OF PERSONS
Milan Adámek, Petr Neumann, Dora Lapková, Martin Pospíšilík and Miroslav Matýsek
Tomas Bata University in Zlín Faculty of Applied Informatics
Algorithms - used for the identification and verification of individuals through fingerprint recognition technology have long been extensively used in Forensic Science and in the private sector. This work is concerned with the verification of the reliability of biometric systems that use fingerprints for their activities. Further, the eFinger programme is used to study similarities between men, women´s and family members´ fingerprints.
INTRODUCTION
Biometrics has been used since ancient times to recognise/distinguish people. People mutually recognised each other by voice, face or the way they walked. Some characteristics do not change during human life; while others, on the contrary continue to be shaped with increasing age [1]. Differentiating people by their fingerprints is one of the oldest Biometric recognition methods. From the earliest times, this method was used by a lot of civilisations that had some form of knowledge of papillary lines, which are included on human skin. The first provable evidence of the use of modern Biometrics however, dates back to somewhere around the mid-19th Century. This was when fingerprints began to be used in Criminology. William James Herschel was one of the first people to take advantage of Biometrics then. He used railway employees´ fingerprints to confirm their identity. Using fingerprints was the only possible way to prove the identity of individual workers, because the majority of them could neither read nor write - and therefore, one could not expect a signature from them. This fingerprint confirmed their identity when being paid their salary. In 1865, Francis Galton came out with a “Study of the Inheritance of Physical Characteristics.” The study dealt with the issue that newly-born babies take over and inherit some characteristics/properties from their parents. These characteristics can include both physical characteristics, as well as some properties - such as, behaviour or conduct. In 1869, Galton became co-
founder of the science called Eugenics, which is the Science of Hereditary Diseases and Defects in the Foetus. A year later, Galton became the founder of research into twins. In 1880, he came up with a branch of science called Anthropometry, which deals with the measurement of human body dimensions. In 1892, Galton published his work entitled "Fingerprints", which led to the introduction of fingerprinting into practice in 1900. In the same year, Galton advocated the use of fingerprinting for identification and verification purposes. He demonstrated the permanence - and uniqueness, of papillary lines on the fingers. After this, fingerprinting/Dactyloscopy was introduced into police work [1]. BIOMETRICS
In Biometrics, several terms exist that are (also) used in Security Technologies. These include identity, identification, authentication, authorisation, verification, and recognition/recognisance. The term Biometrics, is a combination of two words - the word “bio” = life, and “metric” = measurement. Overall then, Biometrics can be seen as a science that deals with the measurement and examination of “live” human characteristics [1] [2]. The notion of identity is derived from the word “idem” - the same. This term is used when - for instance, comparing an object, situation, concept, and such like. One can divide “Identity” into two types; namely, “electronic identity” and “physical identity”. One can have several “Electronic Identities” at the same time – e.g. an identity registered on a Web-site. Conversely, (with regard to) “Physical Identity”, we each have only one, which is unique. Two people, who should have/share the same physical identity, do not exist. It composed of physiological, anatomical and behavioural traits [2][3]. Identification represents the process of discovering and identifying the validity of individuals. To begin with, the person must register such that it passes-on one´s biometric data into the system, which is stored in a database. In the course of determining the identity of a person, a comparison of the information stored in the database (template) and the currently-scanned information (sample) is carried out. This comparison process for as long as it needs to find compliance with data in the database. The output is either - finding the identity … and authorisation to enter; or to refuse entry because there was no consensus in the data.
dy undergoess/characteristiability in timetime - inclu
h almost no cics of the hu- especially
these biomure 1; the degrcentages [4],
Degree of Hussed in percen
SYSTEMS
mponent of ais a sensing metric characaking modules defined in t
+ -
+ -
- +
- +
Identification al Identification al Characteristiceristics tification
s many changics are moree. The two mude the (humchange occursuman voice during puber
metric characgree of tempor[7].
uman Biometntages
a Biometric module that
cteristics. Thele, which cthe database.
● ●
● ●
●
● ●
cs
ges, wherebye – or less,
most consistentman) iris, ands. Conversely,change a lotrty. The timecteristics areral stability is
tric Temporal
Identificationt ensures thee core part iscompares theThe output of
y , t d , t e e s
l
n e s e f
the biomecommunicatithe space pro
Figure 2: Stru[5]
BIOMETRI
One of thesystems inclidentify the officially stodifferentiate/parameters aof the system
• FRRerroused
• FARerroused
Figure 3: DeThreshold, (T
The False AcRate (FRR)occurrence oerrors, it follFAR; and vic(Figure 3.), a“Threshold Vpractice; thasomeone errand FRR the as the EER -determine the
etric identifion interface, ovided.
ucture of a Bi
IC SYSTEMS
e important ludes the abil
identity of ored in a sys/identify anyare used to exm; these are: R – False Rejoneous rejectiod for this is: TR – False Aconeous rejectiod for this is: T
ependence of Th) [4]
cceptance Rat), and expreof a given errows that the hce versa. FRRat the “ThreshValue” dependat is to say, roneously acc values are eq- Equal Error e “approximat
fication syor “lock” - al
iometric Ident
S´ RELIABIL
characteristiclity to clearlythe rightful
stem databasey unknown xpress the deg
jection Rate on) – sometim
Type I Error Rceptance Raton) – sometim
Type II Error R
FAR and FR
te (FAR), andess the proor in percentahigher the FRR and FAR arehold Value”. Tds on the use if the bigge
epts or rejectqual, this equa
Rate. The EEte value of a s
ystem is llowing acces
tification Syst
LITY
cs of biomey and faultles
user - whoe - and also,
persons. Tgree of reliabi
– (probabilitymes, the term aRate te (probability
mes, the term aRate [4][6].
RR on Sensitiv
d False Rejectbability of ages. From thR - the lower e both dependThe setting of
of the systemer problem ists it. When Fality is referredER allows onesecurity system
the s to
tem,
etric ssly
o is , to
Two ility
y of also
y of also
vity
tion the
hese r the dent f the m in s if
FAR d to e to m.”
FIN
A. Opor diftheeva
FigRe
FigTra Ththetheopp B. Thdiffinnumthein t
Fig C. Thdetthidiflin
NGERPRINT
Optical Finge
ptical fingerprtransmission
fferent reflectie space betweealuated throug
gure 4: An Oeflections [3]
gure 5: An ansmission, [3
he optical sense backlightinge nail), and oposite side.
Capacitive Fi
he principle offferences in canger. The senmber of sensoe capacity diffthe papillary r
gure 6: The Pr
Thermal Fing
hermal fingertector” as a hs technology
fference betwees.
T SENSOR P
erprint Sensor
rint sensors arof light. Theseions of light en these linesgh a CCD or C
ptical Sensor
Optical Sens3]
or - using lighg of a finger fon recording t
ingerprint Sen
f this sensor apacity betwensing area isor micro-elect
fference betweridges in a fing
rinciple of a C
gerprint Reade
rprint scanneheat-sensitive is based on m
een peaks and
PRINCIPLES
rs
re based on te sensors expfrom papillar. The reflecte
CMOS sensor
r based on the
sor based on
ht transmissiofrom the uppthe sensor´s i
nsors
is based on meen the sensors equipped wtrodes in ordeeen the peaks
nger.
Capacitor Sens
ers
ers use a selement. The
measuring thed valleys in fin
S
the reflection,loit the use ofry lines - andd light is then.
e principle of
n a Scanning
n, is based onper side (fromimage on the
measuring the-plate and thewith a largeer to evaluates and recesses
sor [3]
small “pyro-e principle ofe temperaturenger papillary
, f d n
f
g
n m e
e e e e s
-f e y
Figure 7: The D. Ultrasonic This sensor transmitter treflected antransmitter oand captured
Figure 8: An FALSE FIN Two approaproduction: 1. Fake finappropriate mcreate a fapreservation characteristic
used in thes expressedto 1000. Thehat is to say,ngerprints. In
re consideredum number ofthan 250, thethe minimum
S8 and S15);S11, S12, S13,nd compared.
ingerprints in
ingerprints in
en comparinge of matching. Despite this,jects does nothe fingerprintmum number
e d e , n d f e
m
; ,
n
n
g g , t t r
Table 4. Cbetween Fam
Furthermore,compliance members. SuSubjects S4,family membyellow colouS5, S6, S12brother, sistetable above, elements of strong matchcase, the matsufficient nucannot be ma CONCLUSI
Biometric swhich is givethis paper wimpair the reexample (pfingerprints –Some typesrecognise impairing tFurthermore,consensus (mfingerprints members. Ththis issue. Itwhen compaset of fingermatches betFingerprints minutia. Thu
ACKNOWL
This work wYouth and SNational Sus(MSMT-777DevelopmenCZ.1.05/2.1.
Comparisons mily Members
, an assessmebetween th
ubjects S2 an, S7 and S15bers - a brothured group in T and S13. Ther, mother - when compfamily mem
h between fatch compliancumber of finatched, or mad
ION
systems are en by the valu
was to suggeseliability of B
presented), is– e.g., by usis of fingerpfingerprint the reliabilit, the study a
matching) – orfor women a
he eFINGER pt is based onaring individurprint compartween familymatches only
us, they can be
LEDGMENT
was supported Sports of thestainability Pr8/2014) and a
nt Fund under 00/03.0089
and Matches
ent was made he fingerpri
nd S10 are sib5 represent aer, sister and Table 4 is madhis quartet isand their co
aring the carmbers, there iamily membece rate is less
ngerprint comde.
closely-linkedues: FAR and st ways that Biometric Sys the produng plastic and
print sensors copies, thu
ty of Biomalso resolves r respectively,and men, andprogramme wn Euclidean Dual points. Evrisons shows y members y on a minim
e distinguished
TS
by the Ministe Czech Repogramme projalso by the Eu
the project C
s in Fingerpr
for matches ints of famblings - brothanother groupcousin. The lde up of Subjes composed oousin). From rdinal fingerpis seemingly
ers. Even in than 250, so
mparison mark
d to reliabilFRR. The aimcan significanstems. One suction of fad rubber mod
are unable us significanmetric Syste
the question, the similarity between fam
was used to tacDistance Metven a very sm
that fingerprare very l
mum numberd from one oth
try of Educatiublic within ject No. LO13
uropean RegioCEBIA-Tech N
rints
and mily hers. p of last, ects of a
the print
no this – a
kers
lity, m of ntly
such false dels.
to ntly
ems. n of y of mily ckle trics mall rint-low. r of her.
ion, the
303 onal No.
RE
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
AU
199of 200theof Zlíareadd
degUnwowaequbetsenresautadd
EFERENCES
RAK, R. Biomcommercial ap247-2365-5. COUFAL, T<http://hw.cz/t
BITTO, O. Etouches.: Comp
LI, Haizhou, Lbiometrics. NeISBN 978-981
JORGENSEN, ZBiometric for Symposium oSecurity. 2011
AHMED, AwadTechnology baDependable an
NAZAR, Akif, IBiometrics for Recognition an
Fujitsu <http://www.fu
UTHOR BIO
MfroCzdeBa
97 to 2008 heTechnology,
08 he has beee Department Applied Inforín, Czech Reeas: power lindress is: adam
Petrthe BTechinduelecR&D
gree in Tecniversity in Zorking in the uas engaged uipment instaltween 1997 anior lecturer asearch work thenticity anadress is: neum
S
metrics and idenpplications, BEN
T. What is teorie-praxe/art20Encryption and puter Media, 200
Liyuan LI a Karew Jersey: Worl-4287-84-5.
Z a T. YU. On MAuthentication.
n Information, ; 476-482
d E. Ahmed a Isased on Mouse Dnd Secure Compussa TRAORÉ a AMouse Dynamic
nd Artificial IntelPalmsecure.
ujitsu.com/cz/solu
GRAPHIES
MILAN ADÁMom the Olomzech Republicgree in Techn
ata Universitye worked as se
Brno Univeren working asof Electronic rmatics of thepublic. Curre
Czech ronics. Havinee in Engineeity in 2013, t the Departmstems of FacBata Universiesearch coverc compatibials for avioncal circuits nsidering theurity issues ara Superior dl. His e-ma
s been graduaUniversity in Zific research
onal defence l securitynical equipmb.cz
K graduated ical UniversityRepublic, ng received ering Informa
he became ment of Compuculty of Applity in Zlín, Czrs the followility, shieldnics, design and testing e security