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

Подождите немного. Документ загружается.

in-focus iris images were used as ground truth (refer-

ence images) in the iris recognition process, as well as

an input to SIRIS for the generation of all conventional

out-of-focus blurred and wavefront coded simulated

imagery. Optical and detector parameters were selected

based on the characteristics of the system used to

collect the ICE images at Notre Dame University

[15]. The F-number was set approximately to 1.92 for

iris images placed at 0.5 m away from the detector. The

pixel pitch is 5.134 mm. The wavefront coded system

was modeled, having a cubic phase element with a ¼30.

The system noise, having a Poisson signal dependent

term associated with the light detection process and a

white Gaussian independent term (1% noise), reﬂect-

ing a camera under ideal lighting and image capture

conditions was modeled.

The blurred images for both conventional and

wavefront coded systems were obtained by modeling

the movement of an iris in 1.13 cm increments

awa y and towar ds the imaging device, in a range

from10.2 cm (away) to 10.2 cm (towards) from

the plane of best focus. This range is signiﬁcantly

wider than the known imaging volume of the standard

system. Figure 3 shows a sample set of blurred images

obtained with a conventional system for a single input

iris image.

Notice that appropriate magniﬁcation of the iris due

to movement away and towards the camera is taken into

account. Figure 4 shows a sample set of blurred images

obtained with a wavefront coded system (a ¼ 30) for

a single input iris image. The wavefront coded images

are unrestored, that is, no digital ﬁlter was applied

to reverse the effects of the cubic phase mask. A total

of 427,500 simulated conventional and wavefront

coded images were generated from the 150 input iris

images.

Wavefront Coded

Iris Biometric Systems. Figure 3 Sample SIRIS simulated standard images.

Wavefront Coded

Iris Biometric Systems. Figure 4 Sample SIRIS unrestored cubic images.

1400

Wavefront Coded

Iris Biometric Systems

The blurred images were then processed using a

research implementation of Daugmans algorithm.

Speciﬁcally, the Hamming distances for all blurred

images with respect to all reference images were

computed. The enrollment images do not contain

defocus blur. For our dataset, the calculation of Ham-

ming distances implies the computation of a total of

2  150  150  19 matching scores (855,000

comparisons). Figure 5 shows the results of the

comprehensive analysis (users 1–50) of imposter and

genuine Hamming distance scores for both the con-

ventional and wavefront coded systems. The increased

operational range for iris recognition evident from

these plots is reported in Table 1.

These results are in accordance with the published

literature [5 , 7] and highlight the efﬁcacy of the simu-

lation tool in generating imagery useful for research in

this ﬁeld. Future work will introduce image restoration

of wavefront coded images in the iris recognition pro-

cess. We expect restored images to produce smaller

Hamming distances and fur ther to increase the opera-

tional range of iris recognition of the wavefront coded

imaging system, at the cost of increasing its computa-

tional requirements. However, the extension in the

operational range of iris bio metric systems using

blurred wavefront coded imagery is an appealing pros-

pect, since it does not increase the computational

requirements of the system and processing time.

Summary

Depth of ﬁeld and operational range play a major

limiting role in the application of iris recognition tech-

nology. The addition of a wavefront coding phase

masks was shown, via simulation, to signiﬁcantly in-

crease the operational range of iris recognition, even in

the absence of restoration schemes. The tradeoff be-

tween degradation in Hamming distance scores and the

inclusion of the cubic phase mask was also demon-

strated. The utility of the simulation software has

been anecdotally veriﬁed against actual blurred and

wavefront coded imagery, an investigation on the use

of wavefront coding imaging technology on a large

image collection is needed. Future initiatives would

consider the development of novel pattern recognition

algorithms to capitalize on the information provided

by a wavefront coded element. The use of segmentation

methods that are less reliant on sharp pupil/iris and iris/

sclera boundaries will also help to improve the perfor-

mance. This paper has shown that wavefront coding

technology can improve the operational range asso-

ciated with iris recognition systems accompanied by a

Wavefront Coded

Iris Biometric Systems. Figure 5 Imposter and genuine score statistics of standard (above) and

unrestored cubic imagery (below) in which the enrollment images are at the plane of best focus.

Wavefront Coded

Iris Biometric Systems. Table 1

Operational range comparison of a conventional versus

a wavefront coded system

System type Distance (cm) Operational range

Conventional 2–2.5 4.5

Wavefront coded 7–5.8 12.8

Wavefront Coded

Iris Biometric Systems

1401

reduction in SNR that does not dramatica lly affect the

performance of the system. Additional work in systems

engineering and performance optimization is necessary

in order to reap the beneﬁts of this technology.

Related Entries

▶ Biometric Sample Acquisition

▶ Biometric System Design

▶ Iris Encoding and Recognition

References

1. Daugman, J.G.: High conﬁdence visual recognition of person by a

test of statistical independence. IEEE Trans. PAMI 15, 1148–1161

(1993)

2. Daugman, J.: How iris recognition works. IEEE Trans. Circuits

Syst. Video Technol. 14(1), 21–30 (2004)

3. Bowyer, K., Hollingsworth, K., Flynn, P.: Image understanding

for iris biometrics: A survey. Tech. rep., University of Notre

Dame, CSE (2007)

4. Narayanswamy, R., Silveira, P.E.X., Setty, H., Pauca, V.P., van der

Gracht, J.: Extended depth-of-ﬁeld iris recognition system for a

workstation environment. In: Proceedings of SPIE Biometric Tech-

nology for Human Identiﬁcation II, vol. 5779. Orlando , FL (2005)

5. van der Gracht, J., Pauca, V.P., Setty, H., Jr., E.R.D., Plemmons,

R.J., Torgersen, T.C., Prasad, S.: Iris recognition with enhanced

depth-of-ﬁeld image acquisition. In: Proceedings of SPIE Defense

and Security Symposium, vol. 5358. SPIE, Orlando, FL (2004)

6. Dowski, E.R., Cathey, W.T.: Wavefront coding for detection and

estimation with a single-lens incoherent optical system. In: Pro-

ceedings of Acoustics, Speech, and Signal Processing, vol. 4,

pp. 2451–2454 (1995)

7. Dowski, E.R., Cormack, R.H., Sarama, S.D.: Wavefront coding:

Jointly optimized optical and digital imaging systems. In: Pro-

ceedings of AeroSense Conference. Orlando, FL (2000)

8. Cathey, W.T., Dowski, E.R.: New paradigm for imaging systems.

Appl. Opt. 41(29), 6080–6092 (2002)

9. Prasad, S., Torgersen, T.C., Pauca, V.P., Plemmons, R.J., van der

Gracht, J.: High-resolution imaging using integrated optical

systems. Int. J. Imaging Syst. Technol. 14(2), 67–74 (2004).

Special Issue: High-Resolution Image Reconstruction I. (invited

paper)

10. Prasad,S.,Pauca,V.,Plemmons,R.,Torgersen,T.,vanderGracht,J.:

Pupil-phase optimization for extended-focus aberration-corrected

imaging systems. In: Proc eedings of SPIE, Advanc ed Signal Pro-

cessing Algorithms, Architectures, and Implementations XIV,

vol. 5559, pp. 335–345. SPIE, Denver, CO (2004)

11. Prasad, S., Torgersen, T., Pauca, V., Plemmons, R., van der

Gracht, J.: Engineering the pupil phase to improve image quality.

In: Z. Rahman, R. Schowengerdt, S. Reichenbach (eds.) Pro-

ceedings of SPIE Visual Information Processing XII, vol. 5108,

pp. 1–12 (2003)

12. Pauca, V., Plemmons, R., Prasad, S., Torgersen, T., van der Gracht,

J.: Integrated optical-digital approaches for enhancing image res-

toration and focus invariance. In: Proceedings of SPIE, Advanced

Signal Processing Algorithms, Architectures, and Implementations

XIII, vol. 5205, pp. 348–357. SPIE, San Diego, CA (2003)

13. Smith, K.N., Pauca, V.P., Ross, A., Torgersen, T.C., King, M.C.:

Extended evaluation of simulated wavefront coding technology in

iris recognition. In: Proceedings of IEEE Conference on Bio-

metrics: Theory, Applications and Systems. SPIE, Washington,

DC (2007)

14. Leonhardt, E., Pauca, V.P.: A User Guide for SIRIS: Simulator of

Iris Recognition Imaging Systems. Wake Forest University, CS

(2004)

15. Phillips, P.J., Scruggs, W.T., OToole, A.J., Flynn, P.J., Bowyer,

K.W., Schott, C.L., Sharpe, M.: Frvt 2006 and ice 2006 large-

scale results. Tech. Rep. NISTIR 7408, National Institute of

Standards and Technology (2007)

Wavefront Coding

Wavefront Coding is a computational imaging tech-

nique that relies on the simultaneou s optimization of

the pupil function and the image processing of a digital

imaging system to increase its tolerance to optical

aberrations. When the aberration being corrected is

defocus, Wavefront Coding is able to provide the sys-

tem designer with an extension of the depth of ﬁeld of

the imaging system.

▶ WavefrontCodingforEnhancingtheImaging

Volume in Iris Recognition

Wavefront Coding for Enhancing the

Imaging Volume in Iris Recognition

PAULO E. X. SILVEIRA,LU GAO,

AMKUMAR NARAYANSWAMY

Omni Vision CDM Optics, Boulder, CO, USA

Synonym

Wavefront Coded

(Wavefront Coded

is a registered

trademark of Omni Vision CDM Optics, Inc.) biomet-

ric iris recognition

1402

Wavefront Coding

Definition

Wavefront Coding™ (Wavefront Coding™ is a trade-

mark of Omni Vision CDM Optics, Inc.) is a compu-

tational imaging technique capable of increasing the

depth of ﬁeld of an imaging system and increasing the

amount of information captured within a given imag-

ing volume. Iris recognition consists of using the

unique texture present in the human iris to perform

biometric identiﬁcation. The small size of the texture

details compared to the subject standoff distance and

normal human motion makes it challenging to capture

high quality iris images using traditional imaging tech-

niques. The use of Wavefront Coding alleviates these

requirements and shows great performance improve-

ments by reducing the constraints in subject position

while maintaining short exposure times during image

capture.

Introduction

The human iris has several unique features that make it

an attractive choice for biometric recognition: (1) the

iris texture is highly unique (possesses high entropy);

(2) the texture is stable over the lifetime of an individ-

ual; and (3) the image of the eye and iris can be

captured from a distance using standard imaging sys-

tems [1–3]. A wide range of imaging systems including

hand-held cameras, miniature cameras, telephoto, and

multi-spectral cameras are currently being considered

for use or are used to capture the image data for iris

recognition. These imaging systems are characterized

by an

▶ imaging volume over which the image of the

iris is captured with adequately high signal quality. The

challenge faced by today’s iris recognition systems is

that subjects must ﬁnd this rather small imaging vol-

ume by trial-and-error. Then, the iris to be recognized

must remain stationary during the image capture time

to avoid motion blur. Thus, the overall subject experi-

ence is not satisfactory and needs improvement.

Ideally, iris capture systems should operate well

with stationary subjects as with moving subjects, and

the image quality should remain high over the entire

imaging volume. This requirement calls for a relatively

fast optical system (low F/#) which, in traditional

imaging systems, inherently leads to a shallow depth

of ﬁeld. On the other hand, an ideal system should

have a large ﬁeld of view and a large dep th of ﬁeld so

that subjects can be id entiﬁed with minimal coope ra-

tion [4, 5]. These seemingly conﬂicting requirements

can be satisﬁed by using

▶ Wavefront Coding,a

computational imaging technique that is capable of

increasing the depth of ﬁeld of an imaging system

without needing to increase its F-number. Note that

the traditional method of stopping down the aperture

to increase the F-number is detrimental to iris recog-

nition in two ways: (1) the brightness of the image,

which is proportional to the signal level, drops propor-

tionally to the area of the obscured aperture; and

(2) the optical resolution drops proportionally to the

reduction in diameter of the aperture.

Wavefront Coded Iris Recognition

System

Wavefront Coding relies on the simultaneous optimi-

zation of the pupil function and the image processing

of a digital imaging system to increase its tolerance to

aberrations that would typically degrade the image [6].

For example, when the aberration being corrected is

defocus, Wavefront Coding is able to provide the sys-

tem designer with an extension of the depth of ﬁeld of

the imaging system [7]. This is done by using non-

conventional asph erical optics to produce point spread

functions (PSFs) that do not vary signiﬁcantly over

an extended imaging volume compared to traditional

optics. Note that the pupil function can be optimized

by inserting a new element at the exit pupil or, more

practically, by altering the shape of existing optical

surfaces. When the goal is to increase the imaging

volume used for biometric iris recognition, Wavefront

Coding can be employed to provide a modulation

transfer function (MTF) of the system that is consis-

tent over the required range of spatial frequencies and

over a wide range of ﬁeld points and defocus [8].

Figure 1 shows the main func tional blocks of a

Wavefront Coded biometric recognition system. Ty pi-

cally, the imaging system consists of a low-F/# digital

imaging system in which the pupil function is opti-

mized to produce PSFs that are mostly invariant with

defocus and MTFs that retain the required information

over a sufﬁciently large range of spatial frequencies.

A digital sensor captures an intermediate (Wavefront

Coded) image and a processed image is generated by

post-processing this image. The processing may be as

simple as a linear ﬁltering step, with ﬁlters that are

Wavefront Coding for Enhancing the Imaging Volume in Iris Recognition

1403

designed in tandem with the Wavefront Coded optics

and have as a goal the conversion of the defocus-

invariant PSFs into diffraction-limited PSFs. The ﬁlter-

ing step consists of a small fraction of the processing

steps required for conventional iris recognition algo-

rithms and, therefore, contributes with minimal pro-

cessing overhead (e.g., see [1–3] for comparison).

The resulting processed image has an extended

depth of ﬁeld and is provided as an input to a biometric

recognition algorithm that uses the same range of spa-

tial frequencies used in the design of the modiﬁed

imaging system including the Wavefront Coded element

and deconvolution ﬁlter. An example of experimental

results obtained using a Wavefront Coded biometric

recognition system for performing robust iris recogni-

tion is detailed below.

Experimental Example

In this example, the performance of a traditional (non-

Wavefront Coded) iris recognition imaging system is

compared with that obtained with a Wavefront Coded

imaging system. In both cases the subject stands at a

distance of about 2.2 m from the imaging system

and the performance is compared in terms of the

range of subject distances over which iris recognition

can be correctly performed. In both cases the image of

the subject’s iris is captured and compared with that

of a pre-enrolled iris, providing a match in terms of a

modiﬁed (normalized) Hamming distance (HD).

The HD is a measure of the number of bits that are

different between any two given

▶ iris codes. A HD of

zero represents a perfect match while a HD of 0.5

is equivalent to randomly varying bits and, therefore,

corresponds to a complete mismatch. In practical

applications, a HD below 0.2 or 0.3 corresponds to a

match between the image of a captured iris and a pre-

enrolled image.

Figure 2 shows a plot of the HD as a function of the

subject position. At each position ten images of two

eyes (a genuine eye and an impostor eye) are captured

in rapid succession. The images are then processed and

compared to a pre-enrolled iris code. All the calculated

HDs for the genuine (impostor) iris are plotte d as

triangles (stars) and the minimum HDs are connected

by a solid (dotted) line. Ten images were collected in

order to provide us with an estimate of the statistical

variation of the HD at each position. In reality, a single

match (a HD below 0.2) would be sufﬁcient to identify

a subject as genuine, making it unnecessary to capture

additional images. Because imperfect image segmenta-

tion is the main cause of variation in the HD, the mini-

mum HDs are selected to aid in isolating this effect when

comparing the performance of the two systems. Note

that in the traditional system the solid line has the

shape of a narrow valley with sharp transition regions,

corresponding to the distances at which defocus causes

a drop in modulation at the spatial frequencies used by

the recognition algorithm. The ﬂat region close to the

best-focus position corresponds to the distances at

which defocus does not impact the spatial frequencies

of interest signiﬁcantly. Beyond this range the image of

the iris becomes so defocused that it is no longer possible

to resolve the details necessary to correctly perform iris

recognition. Using a threshold HD of 0.2, an operating

range of about 16.5 cm is measured for this traditional

imaging system.

Figure 3 shows a plot of the HD as a function of

the iris distance for a Wavefront Coded system with

Wavefront Coding for Enhancing the Imaging Volume in Iris Recognition. Figure 1 Block diagram of a

Wavefront Coded biometric recognition system. The Wavefront Coded imaging system captures an intermediate image,

which is decoded during post-processing. The processed image is then passed along to the biometric recognition

algorithm.

1404

Wavefront Coding for Enhancing the Imaging Volume in Iris Recognition

the same characteristics as that of the traditional imag-

ing system providing the data for Fig. 2. The imaging

system providing the data for Fig. 3 uses the same

detector, same biometric algorithm, and same optical

elements as the one used for the data of Fig. 2, except

for the inclusion of a Wavefront Coded element at the

aperture stop and a decoding step of the intermediate

image before biometric processing. In this case, the

solid line deﬁnes a shallow and broad valley, effectively

demonstrating the trade-off of the lowest HD for an

extended depth of ﬁeld. In this example, Wavefront

Coding provides us with a recognition range of almost

40 cm at a threshold HD of 0.2, providing an extension

of the depth of ﬁeld 2.4 times greater than that which

was achieved with the traditional system while main-

taining a sufﬁciently large discriminatory capacity be-

tween genuine and impostor irises. This extension in the

depth of ﬁeld can be tuned by the system designer and it

translates into greater usability of the iris recognition

system, possibly representing the difference between a

useful system and a system that is so constrained that

has little practical use. In this example, the Wavefront

Coded element was fabricated separately to simplify the

comparison of the traditional test system with the

Wavefront Coded one. In practice, however, it would

be more advantageous to simply modify the shape of

one or more of the existing optical surfaces.

System Design

As me ntioned above, the system designer can tune the

extension of the depth of ﬁeld to match the needs of

the intended application. Therefore a discussion of the

tradeoffs is warranted. The tradeoffs may be well un-

derstood in terms of the signal-to-noise ratio (SNR)

available to the designer, as explained by the HD curve

of a hypothetical biometric recognition system shown

in Fig. 4.

When designing a Wavefront Coded imaging sys-

tem, the system designer is given the choice of how

much to reduce the SNR at best focus in order to

convert it into an extended depth of ﬁeld. As it turns

out, this SNR can be precisely calculated [9] and it is

directly propor tional to the square of the MTF of the

imaging system. Figure 4 shows a plot of the HD as a

Wavefront Coding for Enhancing the Imaging Volume in Iris Recognition. Figure 2 Iris recognition range of a

traditional imaging system. The iris recognition (measured in terms of a modified HD) is plotted as a function of subject

distance showing a narrow recognition range (16.5 cm). Ten measurements are taken at each position and images of a

genuine iris (triangles) are compared to those of an impostor iris (stars). A solid (dotted) line connects the lowest HD of

the genuine (impostor) iris.

Wavefront Coding for Enhancing the Imaging Volume in Iris Recognition

1405

function of the SNR. In general, the HD is a non-linear

but monotonically decreasing function of the SNR.

The binomial distribution of error bits in the iris

provides us with three distinctive regions of interest.

In Region I, the low-SNR region, the SNR is sufﬁ-

ciently low that the HD is approximately constant at

0.5. In Region II, the mid-SNR region, the decrease in

the SNR is proportional to the decrease in the HD. In

Region III, the high-SNR region, the SNR is sufﬁc iently

high that a decrease in the SNR results in a negligible

decrease in the HD. The existence of Region III is

generally possible due to the non-linear processing

used in the generation of the iris codes (e.g., the thresh-

olding step that is typically used to produce the binary

iris code). Region III is clearly the preferred region of

operation, and it demonstrates the fact that it is possi-

ble to extend the depth of ﬁeld of an imaging system

without any noticeable degradation of the HD at best

focus. Then, the job of the system designer consists of

(1) selecting an initial imaging system that operates

preferably in Region III; and (2) simultaneously opti-

mizing the pupil function and the deconvolution ﬁlter

so that the HD is maintained above a threshold.

Wavefront Coding for Enhancing the Imaging Volume in

Iris Recognition. Figure 4 HD as a function of SNR for a

hypothetical biometric recognition system. Note the

existence of three regions of operation: (I) a low SNR

region, where the SNR is so low that identification is

not possible, (II) a mid-SNR region, where a decrease in

the SNR is accompanied by a proportional decrease in the

HD; III) a high SNR region, where a decrease in SNR

produces a negligible decrease in the HD.

Wavefront Coding for Enhancing the Imaging Volume in Iris Recognition. Figure 3 Iris recognition range of a

Wavefront Coded imaging system. The iris recognition (measured in terms of a modified HD) is plotted as a function

of subject distance showing a broad recognition range (40 cm). Ten measurements are taken at each position and

images of a genuine iris (triangles) are compared to those of an impostor iris (stars). A solid (dotted) line

connects the lowest HD of the genuine (impostor) iris.

1406

Wavefront Coding for Enhancing the Imaging Volume in Iris Recognition

It should be noted that this threshold is application-

speciﬁc, and that the optimization should be simulta-

neously performed over the entire imaging volume.

Summary

Wavefront Coding can be used to increase the depth of

ﬁeld of an imaging system by making the system more

tolerant to defocus. This may be done without reducing

the light collection ability of the system or its optical

resolution, as would be the case when using the tradi-

tional method of stopping down the aperture. Moreover,

the additional steps required in image processing are a

small fraction of the total number of steps required for

biometric recognition, translating in an increase in sys-

tem robustness with little overhead. The experimental

example described above demonstrates how Wavefront

Coding can be used to increase the depth of ﬁeld of an

iris recognition system and, by doing so, reduce the

constraints on the subject’s position and effectively in-

crease the usability of the system. Finally, the tradeoff

between the SNR of an imaging system and the extension

of its depth of ﬁeld has been presented and discussed in

association with designing a Wavefront Coded system

for a given biometric application.

Related Entries

▶ Biometric Sample Acquisition

▶ Biometric System Design, Overview

▶ Iris on the Move

▶ Iris Recognition, Overview

References

1. Daugman, J.G.: High conﬁdence visual recognition of person by

a test of statistical independence. IEEE Trans. PAMI 15,

1148–1161 (1993)

2. Daugman, J.G.: The importance of being random: statistical

principles of iris recognition. Patt. Rec. 36, 279–291 (2003)

3. Daugman, J.G.: How iris recognition works. IEEE Trans. Circuits

and Systems for Video Tech. 14(1), 21–30 (2004)

4. Gracht, J. van der, Pauca, V.P., Setty, H., Dowski, E.R., Plemmons,

R.J., Torgersen, T.C., Prasad, S.: Iris recognition with enhanced

depth-of-ﬁeld image acquisition. In: Proceedings of the SPIE

5358, pp. 120–129. Orlando, FL (2004)

5. Plemons, R., Horvath, M., Leonhardt, E., Pauca, P., Prasad, S.,

Robinson, S., Setty, H., Togersen, T., Gracht, J. van der, Dowski, E.,

Narayanswamy, R., Silveira, P.E.X.: Computational Imaging

Systems for Iris Recognition. In: Proceedings of SPIE 5559,

pp. 346–357. Denver, CO (2004)

6. Cathey, W.T., Dowski, E.: A new paradigm for imaging systems.

Appl. Opt. 41, 6080–6092 (2002)

7. Dowski, E., Cathey, W.T.: Extended depth of ﬁeld through wave-

front coding. Appl. Opt. 34, 1859–1866 (1995)

8. Narayanswamy, R., Baron, A.E., Chumachenco, V., Greengard, A.:

Applications of Wavefront Coded Imaging. In: Proceedings of the

SPIE 5299, 163–174. Orlando, FL (2004)

9. Silveira, P.E.X., Narayanswamy, R.: SNR Analysis of Task-based

Imaging Systems with Defocus. Appl. Opt. 45, 2924–2934 (2006)

Wavelength

Electromagnetic radiation includes visible light, ultravi-

olet light, infrared radiation, and radio waves. In classi-

cal electromagnetic ﬁeld theory, electromagnetic

radiation propagates as a pair of coupled electrical and

magnetic ﬁelds. A time varying electrical ﬁeld creates

a time varying magnetic ﬁeld that in turn creates a

time varying electrical ﬁeld that creates a time varying

magnetic ﬁeld that creates a time varying electrical ﬁeld,

and so on. The distance over which the electri-

cal (or magnetic) ﬁeld at a given instant varies from its

maximum to its minimum and back is the wavelength

of the radiation. Interactions between radiation and

matter depend on wavelength. In the human visual

system, wavelength is intimately connected with color.

A wavelength of the order of 440 nm is perceived as blue,

740 nm as red and the human eye is relatively insensitive

to wavelengths longer than 800 nm. On fundamental

physical grounds, the resolution of optical systems is

generally limited to features of the order of the wave-

length of the light used. There are some special cases in

which resolution can be extended beyond these limits.

▶ Biometirc Sample Acquisition

▶ Biometric Sensor and Device, Overview

Wavelet Transform

Wavelet transform divides a given data into different

frequency components at different scales. In biometrics,

Wavelet Transform

1407

operating in wavelet domain provides additional edge

and frequency information useful for fusion and

recognition.

▶ Fusion, Sensor Level

Web Services

The W3C deﬁnes a Web service as ‘‘a software system

designed to support interoperable Machine to Ma-

chine interaction over a network.’’ Web services are

frequently just Web APIs that can be accessed over a

network, such as the Internet, and executed on a re-

mote system hosting the requested services.

▶ Biometric Interfaces

Work-sharing On-card Matching

The reader side executes part of the calculation (such

as alignment) to assist the smartcard to have a faster

biometric veriﬁcation time. The ﬁnal matching score

calculation shall be executed on-card.

▶ On-Card Matching

World Model

▶ Universal Background Models

1408

Web Services

XML Schema Definition

XML Sch ema deﬁnition is a notation (using the XML

syntax) for deﬁning the form of an XML-encoded

message (similar to ASN.1, but with a different syntax

and suppor ting only XML encoding of the data).

▶ Biometric Technical Interface, Standardization

2009 Springer Science+Business Media, LLC