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

Подождите немного. Документ загружается.

Near-Infrared Imaging

Human eyes can only see visible light which occupies

a very narrow band (approximately 400–700 nm wave-

length representing the color range from violet to

red) of the entire electromagnetic spectrum. However,

generally speaking, there is often more information

contained in other bands of the electromagnetic spec-

trum reﬂected from the objects of interest. Some appli-

cations, such as remote sensing of crops, use special

multispectral or hyperspectral imaging instruments to

obtain the object images in a wide spread of bands of

the electromagnetic spectrum. These images show

more detail than is available in the visible light range.

Similarly, while human vein patterns beneath the skin

are invisible under normal visible light conditions they

can be seen using near infrared imaging techniques.

Principle of Near Infrared Imaging

Two special attributes of infrared radiation and human

blood create a different principle for imaging vein pat-

terns: (1) infrared radiation penetrates into biological

tissue to a depth of approximately 3 mm, and (2) the

reduced hemoglobin in venous blood absorbs more of

the incident infrared radiation than the surrounding

tissue [2]. Therefore, by shining an infrared light beam

at the desired body par t, an image can be captured

using a CCD camera with an appropriate infrared ﬁlter

attached to its lens. The location of veins within about

3 mm of the skin surface will appear as darker lines

than the surrounding tissue because less infrared is

reﬂected from where the veins are located due to

it being absorbed in the blood. Biologically, the hemo-

globin has the highest absorption of infrared light in

the range of 800–900 nm [11]. Therefore, the wave-

length of the infrared source should be selected to be

within the near infrared region with wavelength

around 800–900 nm. With this wavelength, it also

avoids undesirable interference from the far infrared

radiation (with a wavelength of 3–14 mm) emitted by

the human body and the environment.

To acquire a near infrared image of the hand, the

hand must be placed on a surface and evenly illumi-

nated by infrared light. The infrared light should emit

peak infrared radiation at about a wavelength of

850 nm. In order to obtain an image with this reﬂected

infrared light from the hand, a CCD camera is needed

whose spectral response also peaks at around 850 nm.

Such cameras are readily available. To reduce the effect

of visible light, an optical infrared ﬁlter of about

800 nm should be mounted on the cameras lens.

Near Infrared Vein Image Quality

Figures 3 and 4 show examples of vein patter n images

captured using a near infrared camera. The veins just

beneath the surface of the skin appear as dark lines.

The near infrared imaging technique can capture the

major vein patterns in the back of the hand as effec-

tively as the far infra red imaging technique as shown

Fig. 3. More importantly, the near infrared technique is

capable of imaging some of the small veins lying in the

palm and wrist areas. Unlike the image of the back of

the hand, where only major veins are visible, the vein

pattern in the palm is far more complex and poten-

tially contains more information than an image of the

back of the hand. This is important because it signiﬁ-

cantly increases the discrimination power of the

vein pattern biometrics when the size of user group is

Hand Veins. Figure 3 Examples of near infrared images of the back of the hand and the underside of the wrist.

690

Hand Veins

large. Near infrared imaging technique is more tolerant

to the external environment and the subject’s body

temperature than far infrared imaging. However, near

infrared images of vein patterns suffers from the dis-

advantage that visible marks on the skin surface are

also visible in the image, which can corrupt the struc-

ture of the vein patterns and lead to problems in the

later imaging processing and pattern recognition

stages. The palm lines are also visible with the vein

patterns as seen in Fig. 4. While human beings are

capable of distinguishing these lines from the vein

patterns in the image, it requires extra effort to remove

these defects using automatic processing of these

images and is particularly difﬁcult if, for example,

there are marks on the palm which are similar in

appearance to veins, and for example when the person

has drawn lines on the hand using a black pen.

A Hand Vein Pattern Matching System

A hand vein recognition system w ill typically consist of

the following processing stages:

1. Hand Vein Image Acquisition

2. Region of Interest Location and Image

Enhancement

3. Vein Pattern Extraction

4. Feature Extraction and Matching Against a Data-

base of Vein Patterns

5. Decision

The methods and problems associated with Stage 1 –

obtaining an image of the veins in the hand have been

described above. Stage 2 – locating the region of inter-

est and enhancing the image – is usually achieved by

extracting the proﬁle of the hand and locating the

valleys between the ﬁngers and the thumb, and using

these as reference points to identify a region (usually a

rectangle) on the palm or back of the hand as shown

in Fig. 5 . Image enhancement is needed because the

clearness of the vein pattern in the extracted region

of interest varies from image to image, therefore, the

quality of these images need to be enhanced before

further processing. There are many pre-processing

techniques available in image processing for image

enhancement. The choice of which to use will depend

on the quality and nature of the image. An example

of region of interest location, extraction, and image

enhancement for a far infrared grey scale image of

the back of a hand is shown in Fig . 6 (a -c).

After image enhancement, processing in Stage 3

is required to extract the vein patterns in the region

of interest. This involves separation of the vein pattern

from the background. Due to the fact that the grey-

level intensity values of the vein vary at different

Hand Veins. Figure 4 Examples of near infrared image of the palm of the hand.

Hand Veins. Figure 5 The typical region of interest in

palm of back of hand vein imaging.

Hand Veins

691

locations in the image, ▶ global thresholding techni-

ques do not usually provide satisfactory results. Hence,

▶ local adaptive thresholding algorithm is usually

required to separate the vein patterns from the back-

ground. The binary image in Fig. 6(d) shows a typical

result of the vein image after thresholding. The vein

image may be processed in this binary form or the

shape of the vein pattern is extracted as a skeleton,

or one pixel wide line image, of the vein path. The

result of a typical skeletonization process is shown

in Fig. 6(e).

Stage 4 of the recognition process involves the

extraction of features from the vein pattern and match-

ing these against the same features extracted from

reference patterns collected from known individuals

and stored in a database of template or reference

vein patterns. This stage remains an area for further

research. The features can be extrac ted from the grey

scale image (Fig. 6(c)), the binary image (Fig. 6(d))or

the skeletonized image (Fig. 6(e)). Previous research

has investigated the matching of vein patterns using

the Hausdorff distance as used in face recognition [12]

and the extraction of minutiae from the skeletonized

image in a similar manner to that frequently applied to

ﬁngerprint image s [13].

The major factor restricting further investigation of

hand veins as a biometric is the lack of a large data-

base of hand vein images for research study. All

reported work carried out to date has involved rela-

tively small databases collected by the individual

researchers. It is therefore not possible to compare

performance results or predict the likely false accep-

tance and false rejection rates that might be expected of

hand vein biometrics.

Summary

The study of hand veins as a biometric has been inves-

tigated sporadically since about 1990. The most suc-

cessful imaging methods use near and far infrared

imaging. Far infrared imaging can capture images of

the large veins in the back of the hand but has difﬁcul-

ties in capturing vein images in the palm because

of the relatively small size of the veins and the resulting

small amount of thermal energy they radiate. Far infra-

red imaging is very sensitive to ambient temperature

and varying human body temperature, which can be

signiﬁcant in an extremity such as the hand. Near

infrared imaging produces good quality images of

Hand Veins. Figure 6 Example of (a) Original far infrared image of the back of the hand, (b) Extraction of the region

of interest, (c) Image enhancement of the region of interest, (d) Extraction of the vein lines, (e) Skeltonization of

the vein lines.

692

Hand Veins

veins just below the surface of the skin as observed

when capturing vein patterns in the back of the hand

or the palm. Near infrared is more tolerant to envi-

ronmental changes because the technique measures

reﬂected infrared and not transmitted infrared. The

major problem with near infrared images is the reten-

tion of visual features such as marks on the skin and

hairs. Detailed study of the processing and matching

of hand vein images needs to be carried out to fully assess

the potential of hand veins as a biometric.

Related Entries

▶ Anatomy of Hand

▶ Finger Geometry, 3D

▶ Hand Databases and Evaluation

▶ Hand Geometry

▶ Hand Shape

▶ Palm Vein

References

1. MacGregor, P., Welford, R.: Veincheck: Imaging for security and

personnel identiﬁcation. Adv. Imaging 6(7), 52–56 (1991)

2. Cross, J.M., Smith, C.L.: Thermographic imaging of subcutane-

ous vascular network of the back of the hand for biometric

identiﬁcation. In: Proceedings of IEEE 29th International Car-

nahan Conference on Security Technology. Sanderstead, IEEE

Publisher, Surrey, UK (1995)

3. Im, S.K., Park, H.M., Kim, S.W., Chung, C.K., Choi, H.S.:

Improved vein pattern extracting algorithm and its implemen-

tation, In: Digest of technical papers of International Conference

on Consumer Electronics, pp. 2–3, IEEE Publisher, Los Angeles,

USA (2000)

4. Wang, L., Leedham, C.G.: A thermal hand vein pattern veriﬁca-

tion system. In: Proceedings of International Conference on

Advances in Pattern Recognition, pp. 58–65, Springer, Bath,

UK (2005)

5. Kumar, A., Prathyusha, K.V.: Personal authentication

using hand vein triangulation. In: Proceedings of the SPIE,

Biometric Technology for Human Identiﬁcation V., vol. 6944,

pp. 69440E–69440E-13 (2008)

6. Lin, C.L., Fan, K.C.: Biometric veriﬁcation using thermal images

of palm-dorsa vein patterns, IEEE Transactions Circuits and

Systems for Video Technology 14(2), 199–213 (2004)

7. Fujitsu-Laboratories-Ltd., Fujitsu laboratories develops technol-

ogy for world’s ﬁrst contactless palm vein pattern biometric

authentication system. Available at http://pr.fujitsu.com/en/

news/2003/03/31.html (2003)

8. Wang, J.-G., Yau W.-Y., Suwandy, A., Sung E.: Person recognition

by fusing palm vein images based on Laplacianpalm representa-

tion. Pattern Recognit. 41(5), 1514–1527 (2008)

9. Miura, N., Nagasaka, A., Miyatake, T.: Feature extraction of

ﬁnger-vein patterns based on repeated line tracking and its

application to personal identiﬁcation. Mach. Vis. Appl. 15,

194–203 (2004)

10. Jain, A., Bolle, R.M., Pankanti, S.: Biometrics: personal iden-

tiﬁcation in networked society. Kluwer Academic Publishers,

Dordrecht (1999)

11. Fujimas, I., Nakazawa, H.: Structural and functional

tissue analysis under skin using near infra-red spectral imaging.

In: Proceedings of the First Joint BMES/EMBS Conference

2, 1114 (1999)

12. Wang, L., Leedham, G.: Infra-red imaging of hand vein patterns

for biometric identiﬁcation purposes. IET Proceedings on

Vision 1(3–4), 113–122 (2007)

13. Wang, L., Leedham, G., Cho, S.Y.: Minutiae feature analysis

for infra-red hand-vein pattern biometrics. Pattern Recognit.

41(3), 920–929 (2008)

Hand Vein Identification

▶ Back-of-Hand Vascular Recognition

Hand Vein Verification

▶ Back-of-Hand Vascular Recognition

Hand-Geometry Device

VITOMIR S

TRUC,NIKOLA PAV E S

Faculty of Electrical Engineering, University of

Ljubljana, Trz

ka 25, SI-1000 Ljubljana, Slovenia

Synonyms

Hand-geometry reader; Hand-geometry scanner;

Hand-scanning device

Definition

Hand-geometry devices are specially designed bio met-

ric devices used for capturing the geometric

Hand-Geometry Device

693

characteristics (e.g., the length, width, thickness and

curvature of the ﬁngers, the palm size, and the dis-

tances between joints ) of a human hand for hand-

geometry-based identity veriﬁcation. A typical hand-

geometry device records images of the lateral and

dorsal parts of the hand using a charge-coupled dev ice

(CCD) camera that is mounted above a ﬂat surface on

which the person presented to the device places his/her

hand. The set of geometrical features extracted from

these images is then matched against a prerecorded

template stored in the device’s database. Depending

on the result of this matching procedure, the identity

of the person presented to the device is either veriﬁed

or not.

Introduction

Hand-geometry devices are among the earliest commer-

cially available biometric devices for automated identity

veriﬁcation [1]. The production and distribution of

the ﬁrst commercial hand-geometry device, called the

Identimat, was launched in the early 1970s by the

Identimation Corp., which adopted a hand-reader

concept developed and patented by Robert P. Miller

[2]. Like with Miller’s original design, Identimation’s

device used spatial characteristics of the index, middle,

ring, and little ﬁngers as the means of establishing the

identity of a person. It utilized a number of electrome-

chanical components and photo electric cells to mea-

sure the length of the four digits and compare them

to ﬁnger-length measurements that were previously

recorded and stored on an identiﬁcation card. The

device was very well received on the market and was

eventually inst alled for access-control purposes at

several high-security facilities run, for example, by

the U.S. Department of Energy, Western Electric, and

U.S. Naval Intelligence [3].

Encouraged by the success of Identimation’s hand-

based veriﬁcation system and the growing demand for

reliable and user-friendly veriﬁcation schemes, several

other companies tried to enter the hand-biometry

market during the 1970s and early 1980s. They devel-

oped numerous prototypes, using ideas and device

designs from early patents (e.g., [2, 4, 5]); however,

most of them never actually made it to the market. One

of the few exceptions was the ‘‘3D hand proﬁle identi-

ﬁcation apparatus’’ devised by David Sidlauskas [6].

His device featured a

▶ platen on which a pers on

placed his/her hand and a digital camera that captured

images of the hand’s side and top views. Discriminative

hand characteristics extracted from these images were

then employed for the identity veriﬁcation. Unlike

previously developed hand-geometry readers, Sidlaus-

kas’ device did not rely solely on two-dimensional

measurements of the hand, but used

▶ orthographic

scanning to capture the hand’s three-dimensional

structure. Later manufactured under the commercial

name ID3D by Recognition Systems Inc. (RSI) [7], it

became an important milestone in the ﬁeld of hand-

geometry-based veriﬁcation and is, albeit in a much

reﬁned form, still on the market today.

By the late 1990s, manufacturers of commercial

hand-geometry devices (such as RSI) were the only dri-

ving force in the development of hand-geometry-based

biometric techn ology. However, in the past decade,

due to advancements in the ﬁelds of biometrics

and computer vision, the academic community has

taken a more active role in the development of hand-

geometry devices, e.g., [1, 8, 9, 10].

Description of the Device

Unlike early hand-geometry devices (e.g., [2, 4]),

which were primarily based on electro-mechanical

components, modern devices, such as the one pre-

sented in Fig. 1, use imaging technology and internal

software to capture and process the images of a per-

son’s hand and to extract the geometrical features, e.g.,

the lengths, widths, thicknesses, and curvatures of the

ﬁngers, and the width, thickness, and area of the palm,

that are used for the identity veriﬁcation.

Hand-Geometry Device. Figure 1 A commercial

hand-geometry device [7].

694

Hand-Geometry Device

The basic design and use of a hand-geometry de-

vice is quite simple. When a person places his/her hand

on the device’s reﬂective, ﬂat surface, referred to as the

platen, he/she ﬁrst has to align his/her ﬁngers w ith a

number of guiding pegs that direct the hand to a

predeﬁned position. The pegs are equipped with pres-

sure sensors which, when enough pressure is applied to

them, simultaneously trigger the charge-coupled de-

vice (CCD) camera (commercial dev ices typically use a

32,000 pixel CCD camera) and the infrared light

source (e.g., light emitting diodes) positioned above

the device’s platen. The platen then reﬂects the emitted

light back to the camera and an image is recorded.

However, as parts of the platen are covered with the

person’s hand, some of the infrared light is absorbed

and only a silhouette is visible in the resulting image

[11, 12, 13]. Because of the design of modern hand-

geometry devices, which feature a side-mounted

mirror inclined at 45

∘

to the platen, the acquired

silhouette image contains both the shap e of the dorsal

(i.e, the top view) as well as the lateral (i.e., the side

view) surfaces of the hand [9]. Once recorded, internal

software extracts a number (more than 90 in commer-

cial devices) of geometrical features from the silhouette

image and uses them to verify the identity of the

person presented to the device.

However, before a person can use the device, he/she

ﬁrst has to enroll. During the enrollment phase, the

device captures several images of the person’s hand,

extracts geometrical features from each of these images

and uses them for the calculation of the template. The

template is then stored in the memory of the device or

on an identiﬁcation card (i.e., a

▶ smart card) and is

later retrieved for comparison. A similar procedure

is required when a person presented to the device is

trying to verify his/her identity. First, the person claims

an identity by entering a personal identiﬁcation num-

ber (PIN) or by swiping an identiﬁcation card

(depending on the input mechanism provided by

the device at hand) through a card-reader module

connected to the device. The device then proceeds

with the image-acquisition and feature-extraction

stages and ﬁnally recalls (either from an internal mem-

ory or from the smart card) the templa te associated

with the claimed identity for comparison. In the

ﬁnal step, a matching procedure is applied to decid e

whether or not the person presented to the device is

who he/she claims to be [7, 9, 11, 12]. In the case of a

positive decision, i.e., the identity of the person is

veriﬁed, the device usually updates the template to

account for possible changes in the geometry of the

person’s hand (which is especially important when the

device is used by children, whose hand-geometry is

changing fast) and stores a new template for future

veriﬁcation attempts in the device’s memory (or ID

card). This process is commonly referred to as tem-

plate averaging [11].

While typical hand-geometry dev ices are designed

to be used in conjunction with the right hand, it is

possible for a person to enroll and verify his/her iden-

tity using the left hand. In this case, the (left) hand is

placed on the platen with the palm facing upwards [11].

As only the geometry of the hand is of signiﬁcance, this

has no negative effect on the veriﬁcation accuracy of

the device.

There are also commercial dev ices available on the

market that do not use the geometry of the whole hand

to verify the identity of a person, but accomplish this

task based on measurements of only two ﬁngers.

The main part of the device is a camera-based

sensor that uses three-dimensional scanning techno-

logy to capture the structure of the index and middle

ﬁnger (of either hand) of the person presented to the

device. From these scans, a set of geometrical features

is extracted and matched against a template recorded

during the enrollment session. Depending on the

outcome of the matching procedure, the identity of

the person presented to the device is either veriﬁed

or not [14].

Research Trends

In recent years, many research groups from private

companies as well as academic institutions have direct-

ed their research towards hand-geometr y-based iden-

tity veriﬁcation. They are developing new veriﬁcation

schemes that require new kinds of hand-geometry

devices, different from those available on the market

today. The main trend at present is to design devices

that require no pegs to control the placement of the

hand. These designs, like the current commercial

devices, still feature a platen upon which the person

places his/her hand. However, as there is no guiding

mechanism, the hand is simply positioned on the

platen with the ﬁngers spread naturally. A CCD camera

or a digital scanner then captures images of the hand

from which pose-independent geometrical features are

extracted and used to verify the identity of the person

presented to the device [10]. Peg-free designs are

Hand-Geometry Device

695

commonly considered to have a number of advantages

over classical hand-geometry devices: ﬁrst, they do not

cause any deformations of the hand’s silhouette shape

as no contact with the guiding pegs occurs; second,

they reduce the chances of the person incorrectly posi-

tioning his/her hand; and third, they allow people with

small hands, e.g., children, who might have problems

reaching all the pressure sensors of current commercial

devices and so are unable to initiate the veriﬁcation

procedure, to use the device. Although numerous ex-

perimental designs following the described trend were

presented in the literature, none of them has yet suc-

ceeded in passing the prototype stage [9, 10].

Characteristics

Over the past 30 years, hand-geometr y-based recogni-

tion has become one of the most popular biometric

technologies for physical access control and time-and-

attendance applications. The broad success of hand-

geometry devices in these speciﬁc application areas

was triggered by various human and operational factors,

among which the following are the most important [9]:

1. User acceptance. Hand-geometry devices offer a

fast, easy to use, and fairly reliable method of user

authentication, and are therefore enjoying a rela-

tively high level of public acceptance. A survey con-

ducted by the Sandia National Laboratories [15]

in 1991 reported that most of the test subjects

favored hand- geometry devices over other devices

based on ﬁngerprint, signature, retina, or face

biometrics.

2. Functionality. Hand-geometry devices can operate

in harsh environmental conditions and are there-

fore suitable for indoor as well as outdoor deploy-

ment [10]. Furthermore, as they rely only on the

geometric structure of the hand, while igno ring its

surface details, they are to some extent insensitive

to the presence of dirt or dust, which makes them a

preferable choice for access control and time-and-

attendance applications in labor-intensive branches

such as the construction industry [7].

3. Template size. The memory requirements for stor-

ing a templa te are the lowest among all biome tric

technologies. With a size of 9 bytes (or 20 bytes for

devices based on the geometry of just two ﬁngers)

they are signiﬁcantly lower than those imposed by

other modalities [11, 12]. Such a small size is

advantageous for three reasons: ﬁrst, when a

hand-geometry device is operating as a stand-

alone unit, it allows a large number of templates

to be stored in the device’s internal memory; sec-

ond, it saves processing time; and third, it permits

the storage of user-templates on identiﬁcation

cards.

4. Failure to enroll (FTE) and failure to acquire (FTA)

rates. Hand-geometry devices can be used by most

of the world’s population, except for some indivi-

duals who suffer from severe arthritic conditions

and are therefore unable to correctly position their

hands on the device’s platen. For this reason, hand-

geometry devices exhibit relatively low FTE and

FTA rates, when compared to other biometric

devices. A recent study involving 200 participants

showed [16] that the FTE and FTA rates for hand-

geometry devices were the lowest among all the

tested devices (tested were face, iris, vein, voice,

hand-geometry, and ﬁngerprint scanners).

Although the presented characteristics resulted in

the widespread use of hand-geometry devices for ac-

cess control and time-and-attendance purposes, there

are, nevertheless, still a number of shortcomings that

limit their use in other application areas, the most

signiﬁcant being:

1. Size. Typical hand-geometry devices are designed to

accommodate the whole human hand (or at least

two ﬁngers) and are consequently signiﬁcantly lar-

ger than other devices used for capturing the bio-

metric traits of a person (e.g., face, voice, and

ﬁngerprints). This fact makes them unsuitable for

security applications, where a compact size for the

biometric device is preferable (e.g., laptops and

mobile devices) [12].

2. Cost. Commercial hand-geometry devices can cost

considerably more than, for example, ﬁngerprint

scanners, which target a similar market segment

(i.e., access contro l and time-and-at tendance).

With a price of approximately $1,000–2,0 00, they

are among the more expensive biometric technolo-

gies [9, 12].

3. Per formance. The false-acceptance (FA) and false-

rejection (FR) rates for hand-geometry-based

security systems are typically higher than those of

ﬁngerprint-, palmprint- or iris-based systems,

which makes hand-geometry devices suitable only

for low/medium security applications. Several

696

Hand-Geometry Device

independent studies (e.g., [13, 15, 16]) reported

that commercial hand-geometry devices achieve

FA and FR rates in the range of 0.1–1.0% at the

equal-error operating point.

Operation

Commercial hand-geom etry devices are capable of

operating in two distinct types of conﬁgurations:

as stand-alone units or as part of a networked system

[7, 11, 17].

1. Stand-alone units. While hand-geometry devices

deployed for time-and-attendance monitoring typ-

ically require an additional time clock and a com-

puter to record and retrieve information about the

arrival and departure times of people, they can,

nevertheless, be used for access control purposes

without the need for any additional components.

A hand-geometr y device can, fo r example, directly

control the locking mechanism o f a door and

release it if the identit y of the person try ing to

gain access to the secured faciliti es is successfully

veriﬁed. In this stand-alone conﬁguration, the

devices are suitable only as access co ntrol systems

for single d oors (e.g., main entrances, doors to

sensitive areas such as computer rooms, storage

areas, etc.), while a (networked) system of hand-

geometry devices is needed when multiple doors

have to be secured. The number of peo ple that can

be enrolled in a stand-alone device is limited by

the storage capabilities of the device, as all user

templates are stored locally in the device’s internal

memory. Furthermore, as no other means are avail-

able, for example, a central computer, all administra-

tive tasks have to be preformed with the help of the

device’s keypad [7, 11, 17].

2. Networked system. Commercial devices support a

number of communication standards and proto-

cols (e.g ., RS-485, RS-422, RS-232, and TCP/IP)

that can be used to connect an a r bitrary num ber o f

devices with a host computer to f orm a networked

system. In contrast to stand-alone units, net-

worked systems are commonly employed in appli-

cations that require multiple hand-geometry

devices (e.g., access control to facilities with seve-

ral entrances). While these requirements can be

met w ith several stand-alone devices, the use of a

networked system h as a number of advantages:

ﬁrst, a person does not have to unde rgo the incon-

venience of enrolling at each of the units, but is

able to enroll at a single location and retrieve his/

her template for comparison from a central stor-

age location at any unit (for which h e/she has

access rights) of the network; second, all door

activiti es and time (and/or attendance) records

can be stored and viewed via a central computer,

making system monitoring simple and efﬁcient;

and third, a networked system enables the cen-

tralized management of user proﬁles (e.g ., their

access rights, and deletions) [7, 17].

Commercial devices are also able to emulate stan-

dard card-reader units, which makes it easy to integrate

them into existing security systems. When employed in

the card-reader emulation mode, the hand-geometry

device, upon successful veriﬁcation, simply forwards

the user’s identiﬁcation number in an appropriate for-

mat to the card-reader module, which then proceeds as

if the identiﬁcation number had been read from an

identiﬁcation card. Several card protocols are com-

monly supported, the main ones being Wiegand, bar-

code, and magnetic stripe [7, 11, 17]. In fact, there are

two standards deﬁning the data-interchange formats

of hand-geometry devices: the ‘‘ANSI INCITS 396-

2005 Hand Geometry Interchange format’’ and its

international counterpart the ‘‘ISO/IEC 19794-10:2007

Biometric Data Interchange Format – Part 10: Hand

Geometry silhouette data.’’ The standards deﬁne both

the format and the content for the exchange of the

hand-silhouette data, and are aimed at increasing

the interoperabilit y of hand-geometry devices [18].

Summary

Among the different biometric devices available on the

market, hand-geometry devices have emerged as the

preferred choice for physical access control and time-

and-attendance applications, especially in harsh envir-

onments where other devices might have problems in

reliably verifying the identity of a person. They are

based on a ﬁeld-proven technology that by today has

been in use for more than 20 years. However, research

is already on the way to produce the next generation of

hand-geometry devices, which will undoubtedly result

in smaller, faster, and more user-friendly units. Several

Hand-Geometry Device

697

research groups have already developed prototypes

that require no guiding pegs to capture an image

suitable for the extraction of hand-geometry features.

While these prototypes still need time to mature, they

are a clear indication of future trends in the develop-

ment of hand-geometry devices.

Related Entries

▶ Anatomy of Hand

▶ Biometric Sensor and Device, Overview

▶ Hand Recognition

References

1. Jain, A.K., Ross, A., Pankanti, S.: A prototype hand geometry-

based veriﬁcation system. In: Proceedings of the Second Inter-

national Conference on Audio- and Video-Based Biometric

Person Authentication (AVBPA), Washington D.C., USA,

pp. 166–171 (1999)

2. Miller, R.P.: Finger dimension comparison identiﬁcation system.

US Patent No. 3576538 (1971)

3. Miller, B.L.: Vital Signs of Identity. IEEE Spectr 31(2), 22–30

(1994)

4. Ernst, R.H.: Hand ID system. US Patent No. 3576537 (1971)

5. Jacoby, I.H., Giordano, A.J., Fioretti, W.H.: Personnel Identiﬁca-

tion Apparatus. US Patent No. 3648240 (1972)

6. Sidlauskas, D.P.: 3D hand proﬁle identiﬁcation apparatus. US

Patent No. 4736203 (1988)

7. Homepage of Recognition Systems Inc.: http://www.recogsys.com

8. Sanches-Reillo, R., Sanches-Avila, C., Gonzales-Marcos, A.:

Biometric identiﬁcation through hand geometry measure-

ments. IEEE Trans. Pattern Anal. Mach. Intell. 22(10),

1168–1171 (2000)

9. Paves

, N., Ribaric

, S., Ribaric

, D.: Personal authentication

using hand-geometry and palmprint features – the state of the

art. In: Proceedings of the Workshop: Biometrics: Challenges

Arising from Theory to Practice, Cambridge, UK, pp. 17–26

(2004)

10. Dutagaci, H., Sankur, B., Yoruk, E.: A comparative analysis of

global hand appearance based person recognition. J. Electron.

Imaging, 17(1), 011018 (2008)

11. Zunkel, R.: Hand geometry based veriﬁcation. In: Jain, A.R.,

Bolle, R., Pankanti, S. (eds.) Biometrics: Personal Identiﬁcation

in Network Society. Kluwer, Dordecht (1999)

12. Homepage of International Biometric Group: http://www.

biometricgroup.com

13. Kukula, E., Elliott, S.: Implementation of Hand Geometry: An

Analysis of User Perspectives and System Performance. IEEE

Aerospace and Electronic Systems Magazine 21(3), 3–9 (2006)

14. Homepage of Biomet Partners Inc.: http://www.biomet.ch/

15. Holmes, J.P., Wright, L.J., Maxwell, R.L.: A Performance Evalua-

tion of Biometric Identiﬁcation Devices. Sandia National

Laboratories, Report, U.S.A. (1991)

16. Mansﬁeld, T., Kelly, G., Chandler, D., Kane, J.: Biometric Prod-

uct Testing: Final Report. Technical Report CESG Contract

X92A/4009309, Centre for Mathematics and Scientiﬁc Comput-

ing, National Physics Laboratory, Middlesex, UK (2001)

17. Spence, B.: Biometrics in Physical Access Control: Issues, Status

and Trends. Available at: http://www.recogsys.com

18. Homepage of the American National Standard Institute: http://

www.ansi.org/

Hand-Geometry Reader

▶ Hand-Geometry Device

Hand-Geometry Scanner

▶ Hand-Geometry Device

Hand-Held Devices

A hand-held device is a pocket-sized computing de-

vice, typically, comprising of a small visual display

screen for user output and a miniature keyboard

or touc h screen for user input. New hand-held

devices include a number of sensors that can be used

to acquire biometric data, e.g., touch-screens (signa-

ture and handwriting), ﬁngerprint sensors, micro-

phones (speech), cameras (face, video), etc.

▶ Fingerprint Databases and Evaluation

Handprint

In a handprint writing style, the writer is inclined to

write each individual character in an isolated fashion.

In other words, there is no connection between adja-

cent l etters.

▶ Signature Sample Synthesis

698

Hand-Geometry Reader

Handprint Sensor

▶ Fingerprint, Palmprint, Handprint and Soleprint

Sensor

Hand-Scanning Device

▶ Hand-Geometry Device

Handwriting Sample Synthesis

▶ Signature Sample Synthesis

Handwriting Synthesis

▶ Signature Sample Synthesis

Handwritten Signature Recognition

▶ Signature Recognition

Head Pose Analysis

▶ Face Pose Analysis

Head Yaw/Tilt/Roll

Yaw, Pitch, and Roll are the three angles of rotation

that describe changes in the orientation of a 3D object,

which in the case of face pose are with respect to the

face being in upright position facing the camera/sen-

sor. Yaw is the left/right rotation angle (head turning),

pitch is the up/down rotation ang le (head nodding)

and roll is the sideways rotation angle (head tilting

while facing the camera/sensor).

▶ Face Pose Analysis

Headspace

Headspace is the gaseous phase above a sample (liquid

or solid ) containing the volatile and semi-volatile com-

pounds released by the substance.

▶ Odor Biometrics

Helper Data

Information extracted from biometric data to assist

aligning or retrieving original biometric template. It

should not leak any information about original tem-

plate but assure alignment accuracy to some extent.

▶ User Interface, System Design

Heterogeneous

Heterogeneous is an adjective used to describe some-

thing that has a large amount of variants or different

forms. Heterogeneity (a noun), resulting from having

different properties, leads to a difference in expected

results, more than can be accounted for by chance.

▶ Heterogeneous Face Biometrics

Heterogeneous

699