Li S.Z., Jain A.K. (eds.) Encyclopedia of Biometrics

Подождите немного. Документ загружается.

The top-lighting imaging method has the following

features.



Robust against environmental illumination.

The light source housing protects the camera

from unwanted ambient lights that deteriorate

the image quality. This structure makes the top-

lighting system the most robust imaging system in

terms of environmental changes.



Stable illumination.

Since the top-lighting system has only one light

source placed right behind the ﬁnger, the contrast

attenuation of the captured image is isotropic and

no special image processing is required as long as the

region of interest has enough signal-to-noise ratio.

Since this is the earliest and the most straightforward

implementation of ﬁnger vein imaging system, many

commercial models today employ this approach for

both logical and physical access control applications.

One of the earliest commercial ﬁnger vein products is a

physical access control system developed by Hitachi

Engineering Co., Ltd. (Its biometrics division was reor-

ganized into Hitachi Information and Control Solu-

tions, Ltd. in 2006 [3].) in 2002. Their product,

SecuaVeinAttestor

employed the top-lighting system

and demonstrated very stable performance. This prod-

uct was further improved in terms of robustness in the

following year and achieved even higher accuracy com-

parable to iris recognition (Fig. 3). Figure 4 shows

a logical access control unit PC-KCA100 jointly devel-

oped by Hitachi, Ltd. (Tokyo, Japan) [ 4] and Hitachi

Software Engineering, Co., Ltd. (Tokyo, Japan) [5]in

2006. This product has an application programming

interface (API) that was developed based upon

▶ BioAPI., which enables the biometric device to easily

communicate with many ty pes of systems. Another

interesting example of top-lighting system was demon-

strated by Hitachi, Ltd. in 2007. It introduced a unique

automobile ignition key device (prototype) in Tokyo

Motor show 2007, which allows pre-enrolled drivers

to start the engine by presenting their ﬁngers on

the ﬁnger vein reader embedded on the steering

wheel [6].

Finger Vein Pattern Imaging. Figure 2 Top-lighting

system (Courtesy of Hitachi, Ltd.).

Finger Vein Pattern Imaging. Figure 3

SecuaVeinAttestor

(Courtesy of Hitachi Information

& Control Solutions, Ltd.).

Finger Vein Pattern Imaging. Figure 4 Hitachi

PC-KCA100 (Courtesy of Hitachi, Ltd.).

430

Finger Vein Pattern Imaging

Side-lighting Systems

Side-lighti ng systems typically have a pair of infrared

LED arrays em bedded on the both sides of the pre-

sented ﬁnger. The infrared rays emitted by the light

source propagate inside the ﬁnger and some of them

reach to the infrared camera placed beneath the ﬁnger

as illustrated in Fig. 5. Figure 6 shows an example of

the side-lighting system.

This imaging method has the following features.



Medium-sized enclosure



User-friendly design; low psychological barrier

Unlike the top-lig hting systems, the presented

ﬁnger is always within the ﬁeld of view of the user,

which considerably reduces psychological difﬁcul-

ties of the user while scanning.



High maintainability

It is easy to clean up the camera surface because no

housing covers the optical unit.

Although the side-lighting systems require very ad-

vanced image processing and illumination control tech-

nologies, it is one of the most popular implementations

that is employed by many commercial models. One of

the most widely used applications of this lighting system

is automated teller machines (ATMs). Hitachi-Omron

Terminal Solutions, Corp. (Tokyo, Japan) [7] is the only

supplier of ﬁnger vein authentication systems for

banking transactions as of 2007, who has shipped

approximately 40,000 ATMs equipped with ﬁnger

vein biometrics (Fig. 7) and enrollment units (Fig. 8)

in Japan since 2005. Hitachi-Omron has also devel-

oped a unique key managem ent system with ﬁnger

vein authentication in 2006 (Fig. 9). Hitachi Software

Engineering, Co., Ltd. developed a compact logic al

access control unit called Johmon J200 in 2004,

which employs the side-lighting system.

Bottom-lighting Systems

Bottom-lighting systems have been developed as an

answer to the growing demand for mobile applica-

tions. Typically, the bottom-lighting systems have a

Finger Vein Pattern Imaging. Figure 6 Infrared LED

array (Courtesy of Hitachi-Omron Terminal Solutions,

Corp.) Infrared LEDs are colored in this picture for

visualization.

Finger Vein Pattern Imaging. Figure 5 Side-lighting

system (Courtesy of Hitachi-Omron Terminal

Solutions, Corp.).

Finger Vein Pattern Imaging. Figure 7 ATM equipped

with a finger vein reader (Courtesy of Hitachi-Omron

Terminal Solutions, Corp.).

Finger Vein Pattern Imaging

431

pair of infrared LED arrays and an infrared camera

embedded on the same surface as shown in Fig. 10.

Although the conﬁguration of the components is

similar to the imaging systems using reﬂectio n lig ht

such as palm vein readers, the ﬁnger has to be touched

to the LED arrays while scanning. The infrared rays

projected into both the tip and the root of the pre-

sented ﬁnger propagate inside the ﬁng er and visualiz e

the vascula r patterns in the same manner as the side-

lighting systems. This imaging method has the follow-

ing features.



Cost effective



Very small in volume

Since both the light source and the camera are

embedded on the same surface, it does not require

any three-dimensional structure. This enables the

imaging system ﬂexibly embedded to many devices

including portable devices such as laptop compu-

ters or mobile phones. The volume of Hitachi’s

prototype unit developed in 2005 [8] is as small

as 19 ml (39 mm (depth)  34 mm (width)  15

mm (height)), and further miniaturization is

expected in the near future.



User-friendly design; minimum psychological

barrier

Since the bottom-lighting systems can be em-

bedded to existing hardware without changing the

original appearance of the hardware signiﬁcantly,

user’s acceptability is the best among the three

systems describe d here. The exterior of the scan-

ning unit is quite similar to the widely used ﬁnger

print scanne rs and thus psychological barrier of the

user is ver y low.



High maintainability

Cleaning the bottom-lighting imaging system is

as easy as wiping a ﬂat surface. In addition, it is not

necessary to clean the system as frequent as other

two systems because it has no holes or ditches in

which dust can accumulate.

Hitachi, Ltd. released the ﬁrst laptop PC equipped

with an embedded ﬁnger vein authentication module

in 2005 by using this imaging technology (Fig. 11).

Finger Vein Pattern Imaging. Figure 8 Hitachi-Omron’s

UBReader (Courtesy of Hitachi-Omron Terminal

Solutions, Corp.).

Finger Vein Pattern Imaging. Figure 9 Key management

system with a finger vein reader (Courtesy of

Hitachi-Omron Terminal Solutions, Corp.) Metal enclosure

is removed for demonstration purpose.

Finger Vein Pattern Imaging. Figure 10 Bottom-lighting

system (Courtesy of Hitachi, Ltd.).

432

Finger Vein Pattern Imaging

Summary

As described in this essay, ﬁnger vein imaging systems

available today can be categorized into three groups:

the top-lighting systems, the bottom-lighting systems

and the side-lighting systems. Because each system has

its unique features, it is very important to choose a

suitable imaging system depending on the application.

In general, the reproducibility of imaging systems is, to

some extent, subject to environmental changes such as

ambient lightings or the conditions of subject, and

of course, none of the three imaging systems over-

viewed here are free from these constraints. In other

words, these changes can be regarded as external

(uncontrollable) parameters and imaging systems

that are robust against these parameters are generally

preferred. The performance of a biometric sy stem can

be improved by suppressing the inﬂuence of these

uncontrollable parameters as little as possible; and it

is particularly important to select an appropriate imag-

ing system depending on the application by taking

the variety and the range of these parameters into

consideration.

Related Entries

▶ Finger Vein

▶ Finger Vein Feature Extraction

▶ Finger Vein Reader

References

1. Hitachi Central Research Laborator y, http://www.hqrd.hitachi.

co.jp/crle/

2. Kono, M., Ueki, H., Umemura, S.: Near-infrared ﬁnger vein

patterns for personal identiﬁcation. Appl.Opt. 41(35),

7429–7436 (2002)

3. Hitachi Information and Control Solutions, Ltd., http://www.

hitachi-ics.co.jp/product/english/index_en.htm

4. Hitachi, Ltd., http://www.hitachi.com/

5. Hitachi Software Engineering, Co., Ltd., http://hitachisoft.jp/

English/index.html

6. Hitachi develops ﬁnger vein authentication technology for steer-

ing wheels. Hitachi, Ltd. News release, Oct. 2007. http://www.

hitachi.com/New/cnews/071022b.html

7. Hitachi-Omron Terminal Solutions, Corp., http://www.hitachi-

omron-ts.com/index.html

8. Hitachi develops compact ﬁnger vein authentication technology

for laptop PCs. Hitachi, Ltd. News release, Oct. 2005. (in Japanese)

http://www.hitachi.co.jp/New/cnews/month/2005/10/1003.pdf

Finger Vein Reader

MIT SUTOSHI HIMAGA

Hitachi-Omron Terminal Solutions, Corp. Tokyo,

Japan

Synonyms

Finger vein scanner; Finger vein imaging device;

Finger vein authentication device

Definition

A ﬁnger vein reader is a biometric device that com-

prises at least one optical imaging unit designed to

capture ﬁnger vein patterns of an individual and a

digital signal processor that digitizes the captured ﬁn-

ger vein patterns to be utilized as biometric features.

Introduction

Unlike conventional biometric features such as ﬁnger

print, vascular network patterns cannot be observed

without using specially designed equipment and

thus, it is extremely difﬁcult to steal or duplicate the

biometric information. The possibility of biometric

Finger Vein Pattern Imaging. Figure 11

Bottom-lighting system (Hitachi Laptop PC Se210)

(Courtesy of Hitachi, Ltd.).

Finger Vein Reader

433

identiﬁcation based on human ﬁnger vein patterns cap-

tured by transmitting light was indicated by Shimizu in

1992 [1]. However, it was not until Kono et al. devel-

oped a near-infrared ﬁnger vein reader prototype and

demonstrated its effectiveness in 2000 that the concept

became reality [2]. Kono et al. further evaluated the

performance of the proposed biometric modality by

using sample data collected from 678 subjects and

reported very positive results in 2002 [3]. In 2007,

Yanagawa et al. demonstrated the diversity of human

ﬁnger vein patterns by conducting statistical analysis

based on sample data collected from 506 subjects. They

also proved the feasibility of using ﬁnger vein patterns

for personal identiﬁcation by evaluating false accep-

tance rates (FAR) and false rejec tion rates (FRR) [4 ].

Today, ﬁnger vein biometrics is recognized as one of

the most reliable and secure biometric modalities and

is applied to a variet y of security systems.

Features of Finger Vein Modality

The advantages of ﬁnger vein biometrics are summar-

ized as below:

1. Accuracy

Finger vein biometrics is one of the most accu-

rate biometric modalities available today. A ﬁnger

vein authentication device called UBReader has

been certiﬁed as level 3 in the accuracy scale by

the US-based International Biome tric Group

[5, 6]. No other biometric device has been rated at

the highest possible level, level 4. The details of the

evaluation results are reported in the Comparative

Biometric Testing (CBT) round 6 Public Report [7].

2. Usability

Finger vein biometrics can be implemented in

many forms according to the demands and require-

ments of the application. This ﬂexibilit y makes it

possible to design the hardware optimized to a spe-

ciﬁc use. For example, Hitachi-Omron’s UBReader,

which was primarily designed for banking applica-

tion, demonstrated very high usability in terms of

indices such as

▶ Failure-to-Enroll Rate (FTE) or

▶ Enrollment Transaction Duration in the CBTand

achieved level 3 in the usability scale of the testing.

3. Compactness/Flexibility

Since the target imaging area of a ﬁnger vein

reader is generally smaller than for other vascular

pattern biometric devices (e.g., palm vein or the

back-of-the-hand vein systems), ﬁnger vein readers

can be installed into a variety of devices ﬂexibly.

One of the most compact ﬁnger vein readers was not

more than 19ml in volume, which made possible for

laptop PCs to embed the device without changing

their appearances. The short focus depth (i.e., the

distance between the camera and the target) makes it

easy to align the ﬁnger and, therefore, no hand-

guide or handle bars, which are sometimes necessar y

for other hand vascular devices, are needed.

4. Small templates

The size of ﬁnger vein template is typically

some hundreds of bytes per ﬁnger. This mean s

that ﬁnger vein biometric database can be very cost-

effective because it does not require a large storage

system, compared with other biometrics. This feature

is also preferable for systems which store templates

on a server and transmit them upon request over a

network. Small template size makes a big difference

especially when a high-speed network is not available

or the data trafﬁc is ver y high.

5. Excellent image quality

Since the raw image is the very ﬁrst input from

which most biometric information is extracted, the

image quality is largely responsible for the overall

performance of the biometric system. All ﬁnger vein

readers, commercially available today, utilize near

infrared rays that are projected through the pre-

sented ﬁnger. The images captured by using this

method (known as the ‘‘light penetration method’’)

have very high contrast and little noise because most

of ridges and wrinkles on the skin are not imaged.

6. More back-up samples

Unlike most biometric systems, ﬁnger vein

biometrics allows more than two templates per

person. Even in the case when one of the enrolled

ﬁngers gets injured and cannot be presented to the

biometric system, it is possible to operate the sys-

tem using other ﬁngers.

The Hardware

Finger vein readers can be classiﬁed into three different

groups, depending on the device used, and execution

of the enrollment and aut hentication processes.

The ﬁnger vein readers in the ﬁrst category do not

have an authentication algorithm on the readers which

434

Finger Vein Reader

is known as ‘‘match-on-PC’’ readers. Instead, the au-

thentication algorithm is implemented as a computer

software and distributed together with the ﬁnger vein

reader. The software is installed to the host PC before-

hand, where the enrollment and authentication pro-

cesses are executed. Since the match-on-PC ﬁnger vein

readers do not need a powerful processor, the cost of

the hardware is relatively low compared with the other

two kinds of ﬁnger vein readers. Due to the low power

consumption, most of the match-on-PC devices can be

driven by the 5 volts power supplied through the uni-

versal serial bus (USB) interface, which contributes to

the compactness and the portability of the device. Since

the turn-around time of the authentication process is

dependent on the host PC’s CPU power and the com-

munication speed of the interface, the throughput of the

entire system may vary. Although the match-on-PC

readers are widely used for the purpose of logical access

control (e.g., PC log-in), they are increasingly coming

into use for physical access control applications.

The second category is called ‘‘match-on-device’’

ﬁnger vein readers. The matc h-on-device reader is

equipped with a CPU that executes both enrollment

and authentication processes inside the reader itself.

The authentication algorithm is implemented in ﬁrm-

ware and is typically encrypted when stored on a non-

volatile static memory. One of the biggest advantages of

this system is that all algorithms and data required for

biometric authentication are enclosed in a

▶ tamper-

proof casing and completely separated from the out-

side world. Since all biometric data and algorithms can

be stored inside of the ﬁnger vein reader, the risk of

hacking is minimal. Another advantage of this system

is that the match-on-device ﬁnger vein readers do not

require high-performance host PCs. In most cases, a

low-performance CPU is enough to communicate and

control the match-on-device ﬁnger vein reader, which

makes it possible to integrate cost-effective systems.

The data communications between the host PC and

the ﬁnger vein reader are limited because no biometric

data is needed to transfer and therefore no high-speed

interface/network is required. The unit price of these

readers tends to be higher than the match-on-PC read-

ers; however, the match-on-device readers can be used

for a w ide range of applications as they are suitable

for both high-securit y systems and low-cost systems.

Typical applications of the match-on-device readers

include banking systems and physical access control

systems.

The third category is known as

▶ match-on-card

ﬁnger vein readers. The authentication algorithm is

implemented as smart card application software and

securely stored onto a smart card together with bio-

metric templates. Upon the host PC’s request, the

match-on-card ﬁnger vein reader extracts the biomet-

ric feature of the presented ﬁnger and sends an authen-

tication command to the smart card together with the

features. The smart card then executes the authentica-

tion algorithm on its own CPU embedded inside and

evaluates the features transmitted by the ﬁnger vein

reader. After the smart card determines whether the

presented ﬁnger matches with the pre-enrolled tem-

plate, it transmits a response back to the host PC

through the reader. One of the beneﬁts of using the

match-on-card system is its high security feature. Both

the authentication algorithm and the template data are

securely stored on a smart card that is inaccessible

without taking validation procedures using Secure Ap-

plication Module (SAM). Since these data is never

transmitted outside the card, the risk of template du-

plication is extremely low. From a viewpoint of system

administration, the risk management cost of the

match-on-card system can be dram atically suppressed

because the system does not need to provide protec-

tion for the template data (the card holders are respon-

sible for their own templates, instead). Though the

authentication processing time is slig htly longer than

other two kinds of readers (this is because the smart

card CPUs are slower than the embedded CPUs or

PCs), it does not make much difference especially for

its primary usage, veriﬁcation. For these reasons,

match-on-card ﬁnger vein systems are currently the

most popular biometric banking solution in Japan.

Security Features

Some ﬁnger vein readers have a security measure called

▶ liveness detection. It is very important for biometric

systems in general to ensure that the enrolled biometric

patterns are genuine. If a biometric device accepts any

artifact mistakenly and enroll it as a genuine template,

that can be used just like a normal key that can be used

by anyone; if this happens, the security level of the

biometric system becomes no higher than convention-

al keys and locks. In the actual applications, enrollment

procedures typically require an administrator to be

present (who will never allow users to enroll artifacts);

Finger Vein Reader

435

however, it is still beneﬁcial to have this security measure

because it is also used in the authentication procedure in

order to ensure that the presented sample is from a live

body. Liveness detection can be implemented by either

hardware or software (or both) and there are many

different methods to realize the functionality. The details

of the method employed by ﬁnger vein readers are,

however, not publicly available at the time of writing

due to the secure nature of the functionality.

Another security feature that some ﬁnger vein read-

ers have is the tamper-proof structure. This structure

enables the system to identify that it has been tampered

with, and in some cases, to disable itself when unautho-

rized person try to dismantle or reverse-engineer the

system. This security measure is especially important

when the biometric system is to be used by open public,

for instance, ATMs. Just like liveness detection, the

details of the tamper-proof structure are highly conﬁ-

dential and no ﬁnger vein manufacturer discloses the

mechanism for security reasons.

Applications



Banking transactions

Banking applications are currently the most

popular application of ﬁnger vein biometrics. The

ﬁrst ﬁnger vein biometric ATM system was devel-

oped and introduced by Hitachi-Omron Terminal

Solutions, Corp. in 2005. The biometric ATM was

equipped with an open-scanning ﬁnger vein reader,

as shown in Fig. 1, and adopted by one of the largest

banks in Japan, Sumitomo Mitsui Banking Corpo-

ration (SMBC, Tokyo, Japan) [8] and later, was

widely adopted by more than 60 ﬁnancial institu-

tions in Japan including Japan Post Bank Co., Ltd.

[9, 10, 11]. According to a recent survey more than

80% of Japanese ﬁnancial institutions that adopted

biometric banking systems employ ﬁnger vein

biometrics [12] . It is expected that more than

40,000 ATMs in Japan will be equipped with ﬁnger

vein readers by the end of 2008, which will make up

approximately 25% of ATMs of the country.

In t ypical ﬁnger vein banking systems, each

account holder who wishes to have his or her bio-

metric data enrolled visits a branch of the bank in

person and enrolls two ﬁngers at the teller counter

after prescribed personal identiﬁcation procedure.

The templates are then stored in a smart card issued

by the bank, on which the matching process is

executed during the authentication process (i.e.,

‘‘match-on-card’’ technology). Since the matching

process is executed against the two templates stored

on the smart card, users can present either of the

two enrolled ﬁngers. Many of the ﬁnger vein ATM

networks are connected to each other and the

account holders can use their biometric bankcards

at any ATM that belongs to the participating ﬁnan-

cial institutions.



Door access control

Door access control is another popular applica-

tion of the ﬁnger vein biometrics. The ﬁrst com-

mercial application of the ﬁnger vein biometrics

was a door access control system called SecuaVei-

nAttestor

developed by Hitachi Engineering Co.,

Ltd. in 2002. (Please note that Hitachi Engineering

Co., Ltd. reorganized its biometrics division into

Hitachi Information and Control Solu tions [13],

Ltd. in 2006.) The door access control system is

equipped with a ten-key pad, with whi ch users type

his or her ID number so that it can execute one-to-

one matching (veriﬁcation). It can also be used

with proximity cards, which allow users to unlock

the door without typing their ID numbers. In

addition, a biometric door access control system

has been developed that works with electric locks

[14, 15]. A prototype automobile entry system

using ﬁnger vein bio metrics was demonstrated

in the Tokyo Motor Show in 2005, which enables

pre-registered users unlock the door just by holding

the door handle (Fig. 2).

Finger Vein Reader. Figure 1 Finger vein reader

implemented on an ATM.

436

Finger Vein Reader



Logical access control

Logical access control is also a popular and

widely used application. Since host computers

(PCs) are normally equipped with a CPU powerful

enough to execute the matching process in real

time, many ﬁnger vein readers for this application

employ the match-on-PC architecture. Hitachi,

Ltd. and Hitachi Software Engineering (Tokyo,

Japan) [16] jointly developed a very compact ﬁnger

vein reader for PC called PC-KCA100 in 2006. This

match-on-PC ﬁnger vein reader has an applicati on

programming interface (API) based on the widely

recognized international standard BioAPI 2.0,

which enables it to easily communicate with many

types of systems. The power consumption of PC-

KCA100 is so small (less than 2.5 watts) that it can

be driven by the power supplied by the USB inter-

face only.



Other applications

Amano Corporation (Kanagawa, Japan) [17]

developed the ﬁrst ‘‘time and attendance’’ termi-

nal equipped with a ﬁnger vein reader called

AGX250AV in 2007. This innovative termin al can

store up to 1000 ﬁnger vein templates and authen-

ticate the users wi thout using an ID card. In addi-

tion to the convenience, AGX250AV eliminates

inappropriate attendance records by impostors

(this is known as ‘‘buddy pu nching’’), which dra-

matically increases the reliability of the time in-

fo rmation system. Alpha Locker System Co., Ltd.

(Kanagawa, Japan) [18] developed the ﬁrst ﬁnger

vein biometric locker FB-BM in 2007. The biomet-

ric locker, which is aimed for public use, has some

tens of doors that can be accessible by presenting a

ﬁnger. FB-BM is capable of identifying a pre-

enrolled ﬁnger by using one-to-many matching

algorithm and does not require the users to specify

which door to open before presenting their ﬁngers.

Summary

Although ﬁnger vein biometrics is one of the latest

biometric technologies, it has already established both

technical and statistical feasibilit y. Finger vein readers

have been successfully applied to a growing array of

applications such as time and attendance or physical

access control systems. Its banking applications remain

one of the largest and the most successful set of appli-

cations for this state-of-the-art biometric modality;

and it is anticipated that more than a quarter of

ATMs in Japan will be equipped with ﬁnger vein read-

ers by the end of 2008. Some ﬁnancial institutions who

adopted other hand vascular biometrics started to

modify their systems to accept ﬁnger vein biometric

data, or even replace their systems with ﬁnger vein

readers. This trend is expected to continue as the

number of ﬁnger vein readers increase, and it is very

likely for the biometric technology to set a new stan-

dard in security applications in the very near future.

Related Entries

▶ Finger Vein

▶ Finger Vein Feature Extraction

▶ Finger Vein Imaging

References

1. Shimizu, K.: Optical trans-body imaging: feasibility of optical

CT and functional imaging of living body. Jpn. J. Medicina

philosophica 11, 620–629 (1992) (in Japanese)

2. Kono, M., Ueki, H., Umemura, S.: A new method for the

identiﬁcation of individuals by using vein pattern matching of

a ﬁnger. In: Proceedings of the Fifth Symposium on Pattern

Measurement, pp. 9–12, Yamaguchi, Japan. (2000) (in Japanese)

3. Kono, M., Ueki, H., Umemura, S.: Near-infrared ﬁnger vein

patterns for personal identiﬁcation. Appl. Opt. 41(35),

7429–7436 (2002)

Finger Vein Reader. Figure 2 Finger vein reader embedded on a door handle.

Finger Vein Reader

437

4. Yanagawa, T., Aoki, S., Ohyama, T.: Human ﬁnger vein images

are diverse and its patterns are useful for personal identiﬁcation.

MHF Preprint Series, MHF 2007–12, Kyushu University 21st

Century COE Program, Development of Dynamic Mathematics

with High Functionality (2007). http://www2.math.kyushu-u.

ac.jp/coe/report/pdf/2007-12.pdf

5. UB R eader is developed by Hitachi-Omr on Terminal Solutions,

Corp. http://www.hitachi-omron-ts.com

6. International Biometric Group, LLC. http://www.biometricgr oup .

com

7. Theme, M., ed.: Comparative Biometric Testing Round 6 Public

Report. The International Biometric Group (2006)

8. Sumitomo Mitsui Banking Corporation, http://www.smbc.co.

jp/global/index.html

9. Japan Post Bank Co., Ltd. http://www.jp-bank.japanpost.jp/

en_index.html

10. Mizuho Bank, Ltd. http://www.mizuhobank.co.jp/english/

11. Reso na Bank, Limited. http://www.resona-gr.co.jp/holdings/

english/index.html

12. Conducted by Hitachi-Omron, January 2008

13. Hitachi Information and Control Solutions, Ltd., http://www.

hitachi-ics.co.jp/product/english/index_en.htm

14. Electric locks developed by Miwa Lock Co., Ltd. http://www.

miwalock.com/

15. Hitachi, Ltd. http://www.hitachi.com/

16. Hitachi Software Engineering, Co., Ltd.http://hitachisoft.jp/

English/index.html

17. Amano Corporation, http://www.amano.co.jp/English/index.html

18. Alpha Locker System Co., Ltd., http://www.alpha-locker.com/

index.html

Finger Vein Scanner

▶ Finger Vein Reader

Fingermark Identification Procedure

▶ Fingerprint, Forensic Evidence of

Fingerprint

Fingerprint is an impression or image left on a surface

by the friction skin of a ﬁnger.

▶ Anatomy of Friction Ridge Skin

Fingerprint Analysis

▶ Fingerprint Features

Fingerprint Authentication

▶ Fingerprint Indexing

Fingerprint Benchmark

▶ Fingerprint Databases and Evaluation

Fingerprint Binarization

Fingerprint binarization is the process of conver ting an

8-bit gray-scale ﬁngerprint image into a 1-bit ridge

image. This is virtually equivalent to thresholding.

Post-processing for the binarized image, such as

smoothing , is also important.

▶ Fingerprint Image Enhancement

Fingerprint Biometric

▶ Fingerprint Recognition, Overv iew

Fingerprint Capture

▶ Biometric Sample Acquisition

438

Finger Vein Scanner

Fingerprint Characteristics

▶ Fingerprint Features

Fingerprint Classification

XUDONG JIANG

Nanyang Technological University, Nanyang Link,

Singapore

Synonyms

Fingerprint indexing; Fingerprint pre-matching; Finger-

print retrieval

Definition

Fingerprint classiﬁcation is a procedure in which ﬁn-

gerprints are grouped in a consistent and reliable way,

such that different impressions of a same ﬁnger fall

into a same group. It can be viewed as a coarse-level

pre-matching procedure so that a query ﬁngerprint

needs to be further compared with only a smaller

subset of ﬁngerprints in the database belonging to the

same group. It is often necessary to integrate a classiﬁ-

cation module into a ﬁngerprint identiﬁcation system

to speed up the database search. A database can be

partitioned into

▶ human-interpretable ﬁngerprint

classes based on Galton–Henry scheme or into

▶ machine-generated ﬁngerprint classes.

Introduction

A ﬁngerprint recognition system captures a user’s ﬁn-

gerprint and compares it with the information stored

in a database to establish or to authenticate his/her

identity. If an identity is claimed, the system compares

the query ﬁngerprint only with the template

corresponding to this identity stored in the database.

This one-to-one matching process is called ﬁngerprint

veriﬁcation. If no identity is claimed, the system needs

to compare the query ﬁngerprint with all templates

stored in the database to establish the identity. This

one-to-many matching process is called ﬁngerprint

identiﬁcation. The extension of the one-to-one match-

ing of a veriﬁcation system to the one-to-many match-

ing of an identiﬁcation system increases the possibility

of false positive matching. Comparing to the veriﬁca-

tion performance, both accuracy and speed may deteri-

orate signiﬁcantly if a veriﬁcation algorithm is naively

extended to solve an identiﬁcation problem. The per-

formance deterioration could be very serious for large-

scale identiﬁcation systems as it is directly proportional

to the number of ﬁngerprints in the database [1]. This

problem can be alleviated by reducing the search space

of exact matching. Fingerprint classiﬁcation, indexing,

or retrieval techniques facilitate the reduction of the

search space. They can be viewed as a coarse-level pre-

matching process before further exact matching in an

identiﬁcation system. A query ﬁngerprint is ﬁrst com-

pared to prototypes of the pre-speciﬁed classes, bins or

clusters to ﬁnd its class membership. Then, it is only

necessary to compare the query ﬁngerprint exactly

with a subset of the database that has the same class

membership. For example, if a database is partitioned

into ten groups, and a query ﬁngerprint is matched to

two of the ten prototypes, then the identiﬁcation sys-

tem only needs to search two of the ten groups of the

database for exact matching. This reduces the search

space by ﬁvefold if ﬁngerprints are uniformly

distributed in the ten groups.

The ﬁrst rigorous scientiﬁc study on ﬁngerprint clas-

siﬁcation was made by Sir Francis Galton in the late

1880s [2]. Classiﬁcation was introduced as a means of

indexing ﬁngerprints to speed up the search in a data-

base. Ten years late, Edward Henry reﬁned

Galton’s work and introduced the concept of ﬁnger-

print ‘‘core’’ and ‘‘delta’’ points for ﬁngerprint classiﬁ-

cation [3]. Figure 1 shows the ﬁve most common

classes of the Galton–Henry classiﬁcation scheme

where the core and delta points and the class names are

shown. Henry’s classiﬁcation scheme constitutes the

basis for most modern classiﬁcation schemes. Most law

enforcement agencies worldwide currently employ some

variants of this Galton–Henry classiﬁcation scheme. Al-

though Galton–Henry scheme has some advantages,

such as human-interpretable and rigid segmentation of

a database, only a limited number of classes are applica-

ble to the automated system. For example, most auto-

mated systems [4–8] can only classify ﬁngerprints into

ﬁve classes as shown in Fig. 1. Moreover, ﬁngerprints

Fingerprint Classification

439