Yanushkevich S.N., Wang P.S.P., Gavrilova M.L., Srihari S.N. (eds.) Image Pattern Recognition. Synthesis and Analysis in Biometrics
Подождите немного. Документ загружается.
April 2, 2007 14:42 World Scientific Review Volume - 9in x 6in Main˙WorldSc˙IPR˙SAB
358 Synthesis and Analysis in Biometrics
showed significant differences, so too did:
Speed of the signature (p-value = 0.002),
Speed variation (p-value = 0.006),
Average velocity of the signing (p-value = 0.000),
Up-speed (the speed when the pen is off the digitizer) (p-value = 0.000),
Net speed (p-value = 0.000), and
Average uptime (p-value = 0.000)
were all significantly different.
When signing restriction was removed in
[
11
]
, out of the 91 variables,
53 were significantly different at a =0.01. The signing space on the
Wacom digitizer was 8 × 6 inches, and the signing space on the PDA was
2.3 × 2.3 inches. From this difference, the signing area variables, including
the bounding area of the signature, net area, and the ratio of signature ink
length to area, were all significantly different. This result would validate
the assumption that people sign with a larger signature when given a larger
signing space. Again, the variables that differed on the SPT 1500 compared
to the template signature on the Wacom digitizer were the local maximum
or minimum ratios. Other variables that were statistically significant were
velocity, the time the pen was on the digitizer, the average speed that the
pen was on the digitizer, and the net speed of the signature.
The final hypothesis was to examine whether there were any changes on
one particular digitizer when an individual was sitting and then standing.
There were no restrictions on the signing space. The motivation for this
hypothesis was that in many environments where this digitizer was used,
the user would sign both seated and standing.
Example 17. In many retail environments, customers stand when signing
a credit card receipt at a checkout; but in cases where the Interlink E-Pad is
currently being used, for example in banking and brokerage, the user may
also be seated. No differences were statistically significant when a t-test
was performed at α =0.01, showing that when a member of the volunteer
crew stands and enrolls and then sits to sign, the variables of the signature
do not differ statistically at the α =0.01 level.
Based upon the research, the major conclusion that can be drawn from
the study is that
April 2, 2007 14:42 World Scientific Review Volume - 9in x 6in Main˙WorldSc˙IPR˙SAB
Issues Involving the Human Biometric Sensor Interface 359
There are significant differences in signature trait variables across
devices, yet these variables are not significantly different within device
families.
When these devices are grouped together, these variable differences
continue to be significant. The fact that the signature variables differ across
devices is significant. Dynamic signature verification is used within the
realms of electronic document management and contracts and will typically
be verified only if the document validity is questioned. Therefore when
deploying dynamic signature verification applications, it is important to
notethatthedesignofthedevicecanhaveanimpactonthevariables
collected, and depending on the weighting of the variables within a specific
algorithm, the HBSI may have an impact on the system performance.
Furthermore, for audit control purposes, the type of device needs to be
embedded in some way to the signature so that document examiners can
compare signatures captured on the same type of device. Based upon
the data, the major conclusion that can be drawn from the study is that
there are significant differences in the variables across devices, yet these
variables are not significantly different within device families from Wacom
Technologies, Palm and Interlink Technologies.
14.6. Summary
Ergonomics, environment, biometric sample quality, and device selection
have a direct impact on analysis of biometric data. The interaction of these
factors has a significant impact on the performance of a biometric system,
as shown by the different studies mentioned throughout this chapter. The
research in this chapter indicates that a better understanding of ergonomics,
environmental issues, sample quality, and device interoperability is crucial
if biometric technologies are to become more pervasive. All the issues
mentioned, along with HBSI, are open research topics which need to be
addressed to optimize the analysis of biometric data.
Acknowledgments
This research has been supported by Purdue University Department
of Industrial Technology, College of Technology; the Purdue University
Discovery Park e-Enterprise Center, and the Center for Education,
Research, and Information Assurance and Security at Purdue University.
April 2, 2007 14:42 World Scientific Review Volume - 9in x 6in Main˙WorldSc˙IPR˙SAB
360 Synthesis and Analysis in Biometrics
Bibliography
1. Alim, O., Elboughdadly, N., El Morchedi, C., Safawat, E., Abouelseoud, G.,
El Badry, N., Mohsen, N. (2000). Identity Verification using Audio-Visual
Features. Proc. 17th National Radio Science Conf., Egypt, pp. C12/1–C12/8.
2. American Academy of Dermatology. (2002). Mature Skin. Retrieved Nov. 14,
2002, http://www.aad.org/NR/exeres/
3. Ansano, H., Sumi, A., Ramzan, Z., and Zhu, I. (2000). Wireless Electronic
Commerce Security. Retrieved March 22, 2000, http://wwvv.nokia.com
4. ANSI INCITS. (2004). Information Technology Face Recognition Format for
Data Interchange: American National Standards Institute - International
Committee for Information Technology Standards. No. 385-2004. 66p.
5. Behrens, G. (2002). Assessing the Stability Problems of Biometric Features.
Proc. Intl. Biometrics Conf. and Exhibition,Amsterdam,NL.
6. Bioscrypt. (2003). Bioscrypt’s Pattern-Based Algorithm. Retrieved June 20,
2004, http://www.bioscrypt.com/assets/bioscrypt
algorithm.pdf
7. Blackburn, D., Bone, M. and Phillips, J. (2001). Facial Recognition Vendor
Test 2000 Evaluation Report. 61p.
8. Buettner, D. (2001). A Large-Scale
Biometric Identification System at the Point of Sale. Retrieved Sept. 29,
2002, http://www.itl.nist.gov/div895/isis/bc2001/FINAL
9. Clarke, R. (1994). Human identification in information systems: Management
challenges and public policy issues, Information Technology and People,7,
pp. 6–37
10. Dullink, H., Van Daalen, B., Nijhuis, J., Spaanenburg, L., and Zuidhof,
H. (1995). Implementing a DSP Kernel for Online Dynamic Handwritten
Signature Verification Using the TMS320DSP Family (SPRA304), France:
EFRIE, 26p.
11. Elliott, S. (2001). A comparison of on-line Dynamic Signature Trait Variables
across different computing devices, Ph.D. Thesis, Department of Industrial
Technology, Purdue University, 215p.
12. Elliott, S. (2002). A Comparison of On-Line Dynamic Signature Trait
Variables vis--vis Mobile Computing Devices and Table-Based Digitizers, 3rd
IEEE Workshop Automatic Identification Advanced Technologies, pp. 121–
125
13. Greiss, F. (2000). On-Line Signature Verification, M.S. Thesis, Department
of Computer Science, Michigan State University, 88p.
14. Hamilton, D., Whelan, J., McLaren, A., and MacIntyre, I. (1995). Low Cost
Dynamic Signature Verification System, Proc. of European Convention on
Security and Detention, pp. 202–206
15. Holmes, J., Wright, L., and Maxwell, R. (1991). A Performance Evaluation
of Biometric Identification Devices, Albuquerque: Sandia National Labs No.
SANDIA91-0276.
16. INCITS (2003). Finger Pattern Based Interchange Format (Version 7).
Washington, DC: INCITS.
April 2, 2007 14:42 World Scientific Review Volume - 9in x 6in Main˙WorldSc˙IPR˙SAB
Issues Involving the Human Biometric Sensor Interface 361
17. Interlink. (2001). E-Pad Data Sheet. Retrieved July 4, 2001,
http://www.interlinkelectronics.com
18. International Ergonomics Association (IEA). (2006). The Discipline of
Ergonomics. Retrieved Jan. 27, 2006, http://www.iea.cc/ergonomics/
19. Jain, A., Hong, L., Pankanti, S., and Bolle, R. (1997). An Identity-
Authentication System Using Fingerprints, Proceedings of the IEEE, 85(9),
1365–1388.
20. Jain, L., Halici, U., Hayashi, I., Lee, S., and Tsutsui, S. (Eds.). (1999).
Intelligent Biometric Techniques in Fingerprint and Face Recognition. CRC
Press, Boca Raton, FL.
21. Jiang, X., and Ser, W. (2002). Online Fingerprint Template Improvement.
Proc. Pattern Analysis and Machine Intelligence, 24 (8), pp. 1121–1126.
22. Komiya, Y., and Matsumoto, T. (1999). On-Line Pen Input Signature
Verification PPI. Proc. 6th Int. Conf. Neural Information Processing, 3, pp.
1236–1240.
23. Kukula, E. (2004). The Effects of Varying Illumination Levels on FRS
Algorithm Performance, M.S. Thesis, Department of Industrial Technology,
Purdue University, 98p.
24. Kukula, E. (2004). Effects of Light Direction on the Performance of
Geometrix FaceVision 3D Face Recognition System, Proc. Biometric
Consortium, Retrieved Oct. 21, 2004,
http://www.biometrics.org/bc2004/program.htm.
25. Kukula, E., and Elliott, S. (2004). Effects of Illumination Changes on the
Performance of Geometrix FaceVision 3D FRS, Proc. 38th Annual Int.
Carnahan Conf. Security Technology (ICCST), pp. 331–337.
26. Leclerc, F., and Plamondon, R. (1994). Automatic Signature Verification,
Int. J. of Pattern Recognition and Artificial Intelligence, 8(3), pp. 673–660
27. Lee, H. and Gaensslen R. (2001). Advances in Fingerprint Technology (2nd
ed.) Elsevier, New York.
28. Lee, L., Berger, T., and Aviczer, E. (1996). Reliable On-Line Human
Signature Verification Systems, IEEE Trans. Pattern Analysis and Machine
Intelligence, 18(6), pp. 643–647.
29. Lee, W., Mohankrishnan, N., and Paulik, M. (1998). Improved Segmentation
through Dynamic Time Warping for Signature Verification Using a Neural
Network Classifier, Proc. Int. Conf. Image Processing, pp. 929–933.
30. Li, X., Parizeau, M., and Plamondon, R. (1998). Segmentation and
Reconstruction of On-Line Handwritten Scripts, Pattern Recognition, 31(6),
pp. 675–684.
31. Mansfield, T., Kelly, G., Chandler, D., and Kane, J. (2001).
Biometric Product Testing Final Report. Retrieved March 17, 2005,
www.cesg.gov.uk/site/ast/biometrics/media/BiometricTestReportpt1.pdf
32. Mansfield, A., and Wayman, J. (2002). Best Practices in Testing and
Reporting Performances of Biometric Devices: Biometric Working Group.
32p.
33. Martens, R., and Claesen, L. (1997). On-Line Signature Verification:
Discrimination Emphasised, Proc. 4th Int. Conf. on Document Analysis and
April 2, 2007 14:42 World Scientific Review Volume - 9in x 6in Main˙WorldSc˙IPR˙SAB
362 Synthesis and Analysis in Biometrics
Recognition, pp. 657–660.
34. Mingming, M., and Wijesoma, W. (2000). Automatic On-Line Signature
Verification Based on Multiple Models, Proc. IEEE/IAFE/INFORMS Conf.
on Computational Intelligence for Financial Engineering, pp. 30–33.
35. Narayanaswamy, S., Hu, J., and Kashi, R. (1999). User Interface for a PCS
Smart Phone, Proc. IEEE Int. Conf. Multimedia Computing and Systems,
pp. 777–781.
36. Phillips, P., Grother, P., Bone, M., Micheals, R., Blackburn, D., and Tabassi,
E. (2003). Face Recognition Vendor Test 2002 Evalulation Report. 56p.
37. Phillips, P., Martin, A., Przybocki, M., and Wilson, C. (2000). An
Introduction to Evaluating Biometric Systems, Computer, 33, pp. 56–63.
38. Phillips, P., Rauss, P., and Der, S. (1996). FERET (Face Recognition
Technology) Recognition Algorithm Development and Test Report. 73p.
39. Plamondon, R., and Parizeau, M. (1988). Signature Verification from
Position, Velocity and Acceleration Signals: A Comparative Study, Proc.
9th Int. Conf. Pattern Recognition, pp. 260–265.
40. Podio, F. (2002). Personal Authentication Through
Biometric Technologies.Proc. 4th Int. Workshop on Networked Appliances,
pp. 57–66.
41. Rood, E., and Jain, A. (2003). Biometric Research Agenda: Report of the
NSF Workshop. 17p.
42. Sanderson, S., and Erbetta, J. (2000). Authentication for Secure
Environments Based On Iris Scanning Technology, Proc. IEEE Colloquium
Visual Biometrics, 8, pp. 1-7.
43. Schmidt, C., and Kraiss, K. (1997). Establishment of Personalized Templates
for Automatic Signature Verification, Proc. 4th Int. Conf. Document Analysis
and Recognition, pp. 263–267.
44. Sickler, N. (2003). An Evaluation of Fingerprint Quality across an Elderly
Population vis--vis 18-25 Year Olds, M.S. Thesis, Department of Industrial
Technology, Purdue University.
45. Sims, D. (1994). Biometric Recognition: Our Hands, Eyes, and Faces
Give Us Away, Proc. IEEE Computer Graphics and Applications, 14(5),
pp. 14–15.
46. Starkey, R., and Aleksander, I. (1990). Facial Recognition for Police Purposes
Using Computer Graphics and Neural Network, Proc. IEEE Colloquium on
Electronic Images and Image Processing in Security and Forensic Science,
pp. 1–2.
47. Sutherland, K., Renshaw, D., and Denyer, P. (1992). Automatic Face
Recognition, Proc. 1st Int. Conf. Intelligent Systems Engineering, pp. 29–
34.
48. Tayyari, F., and Smith, J. (2003). Occupational Ergonomics: Principles and
Applications. Kluwer, Norwell.
49. Wacom Technology Co. (2000). Wacom Product Specification. Retrieved
November 20, 2000, from http://www.wacom.com
50. Yamazaki, Y., Mizutani, Y., and Komatsu, N. (1999). Extraction of
Personal Features from Stroke Shape, Writing Pressure, and Pen Inclination
April 2, 2007 14:42 World Scientific Review Volume - 9in x 6in Main˙WorldSc˙IPR˙SAB
Issues Involving the Human Biometric Sensor Interface 363
in Ordinary Characters. Proc. 5th Int. Conf. Document Analysis and
Recognition, pp. 426–429.
51. Yanushkevich, S.N., Stoica, A. and Shmerko, V. P. (2006). Fundamentals of
Biometric-Based Training System Design, this issue.
April 2, 2007 14:42 World Scientific Review Volume - 9in x 6in Main˙WorldSc˙IPR˙SAB
This page intentionally left blankThis page intentionally left blank
April 2, 2007 14:42 World Scientific Review Volume - 9in x 6in Main˙WorldSc˙IPR˙SAB
Chapter 15
Fundamentals of Biometric-Based Training System Design
S. N. Yanushkevich
∗
,A.Stoica
†
,V.P.Shmerko
‡
Biometric Technology Laboratory: Modeling and Simulation,
University of Calgary, Canada
Humanoid Robotics Laboratory, California Institute of Technology,
NASA Jet Propulsion Laboratory, California Institute of Technology,
Pasadena, CA, USA, and ECSIS - European Center for Secure
Information and Systems, Iasi, Romania
Biometric Technology Laboratory: Modeling and Simulation,
University of Calgary, Canada
∗
syanshk@ucalgary.ca
†
adrian.stoica@jpl.nasa.gov
‡
vshmerko@ucalgary.ca
This chapter describes the fundamental concepts of biometric-based
training system design and training methodology for users of biometric-
based access control systems. Such systems utilizes the information
provided by various measurements of an individual’s biometrics. The
goal of training is to develop the user’s decision making skills based on
two types of information about the individual: biometric information
collected during pre-screening or surveillance, and information collected
during the authorization check itself. The training system directly
benefits users of security systems covering a broad spectrum of social
activities, including airport and seaport surveillance and control,
immigration service, border control, important public events, hospitals,
banking, etc.
Contents
15.1. Introduction ................................... 366
15.2. BasicsofTrainingSecuritySystemDesign .................. 370
15.2.1. Requirements of Training Systems . . . . . . . . . . . . . . . . . . 371
15.2.2. Model of the Biometric-Based Access Control . . . . . . . . . . . 373
15.2.3. Biometrics............................... 376
15.2.4. Cooperation Between the Officer and the Screened Individual . . 379
365
April 2, 2007 14:42 World Scientific Review Volume - 9in x 6in Main˙WorldSc˙IPR˙SAB
366 Synthesis and Analysis in Biometrics
15.2.5. Decision Making Support . . . . . . . . . . . . . . . . . . . . . . . 382
15.3. TrainingEstimationTechniques ........................ 385
15.3.1. BayesianStatistics .......................... 386
15.3.2. DegradationFunction......................... 387
15.3.3. TrainingFunction........................... 387
15.3.4. EvolutionFunction .......................... 388
15.4. TrainingSystemDesign............................. 390
15.4.1. DesiredCharacteristics........................ 390
15.4.2. Architecture.............................. 391
15.4.3. Training Scenarios . . . . . . . . . . . . . . . . . . . . . . . . . . . 393
15.4.4. Decision Making Support . . . . . . . . . . . . . . . . . . . . . . . 394
15.5. TrainingSystemPrototype........................... 395
15.5.1. Platform................................ 395
15.5.2. DataFlow ............................... 395
15.5.3. Indirect Characteristic Computing . . . . . . . . . . . . . . . . . . 396
15.5.4. ConversionofData .......................... 397
15.5.5. Modularization ............................ 398
15.5.6. Estimations .............................. 398
15.6. DiscussionandOpenProblems ........................ 399
15.6.1. NovelApproaches........................... 400
15.6.2. OpenProblems ............................ 400
15.7. Conclusion.................................... 402
Bibliography ....................................... 403
Glossary
BBN — Bayesian Belief Network
HQP — High Quality Personnel
IR — Infrared Image (band, spectrum)
PASS — Physical Access Security System
SPA — Security Personnel Assistant
T-PASS — Training Physical Access Security System
15.1. Introduction
This paper presents a summary of theoretical results and design experience
obtained in investigations conducted at the Biometric Technology Labora-
tory, University of Calgary, Canada, in collaboration with the Humanoid
Robotics Laboratory at the NASA Jet Propulsion Laboratory, California
Institute of Technology, in Pasadena, California
[
52
]
. These investigations
focus on using biometrics as early warning information rather than data for
authentication. Initial studies led to the idea of designing and prototyping
April 2, 2007 14:42 World Scientific Review Volume - 9in x 6in Main˙WorldSc˙IPR˙SAB
Fundamentals of Biometric-Based Training System Design 367
a biometric-based Physical Access Security System (PASS) using an early
warning paradigm and integrated knowledge-intensive tools at the global
level with distributed functions.
Many research and industrial applications of training systems exist in
such diverse domains as flight simulation
[
13,24
]
,drivingsimulation
[
2
]
,
environment simulation
[
30
]
, and surgical simulation
[
27
]
.
The military has long experience in simulation and training. The rapid
pace of development of new military hardware requires simulators that are
flexible, upgradeable, and less expensive. For example, there is a need
to replace aircraft components on flight line trainers with simulated units
designed specifically for this application.
Another trend is remote simulation through networking, eliminating the
need to transport trainees to the simulator site. Networking is also requires
in team simulations
[
26
]
.
Surgical simulation systems have been developed for training medical
personnel and for testing new technologies which might prove unreliable
for direct use in a real clinical intervention. Image-guided surgery systems
use virtual surgical tools in a 3D virtual scene. These systems provide
the surgeon with a spatial image formation which he/she previously would
have had to reconstruct mentally. Finally, robotic surgery strives for the
minimization of human error during surgery.
Effective utilization of all available information, and the intelligence to
process it, is a key objective of a biometric-based PASS
[
51,52
]
. Binding all
sources of information into this objective can provide reliable identification
of an individual. A Guidance Package Biometrics for Airport Access
Control
[
4
]
provides criteria that airport operators need to know in order to
integrate biometric devices into their access control systems and to revise
their security plans. However, PASS is a semi-automated system; that is,
the final decision must be made by the highly qualified personnel (HQP)
human users. The HQP must be trained to use these tools in practice.
The fundamental premise of our approach is that the training of HQP can
be carried out on the existing PASS itself, expanded with knowledge-based
analytical and modeling tools.
Two directions in access control system design can be discern today
[
1,8,45
]
. The first direction includes approaches that are based on the
expansion of sources of biometric data and, as a result, the amount of
professional skills required of the officer increases. In these approaches, the
mathematical problems of supporting the officer are critically simplified,
allowing a cost-efficient, but in practice not very suitable, solution. The
second direction aims at high-level automated system design, where the
human factor is critically reduced. Such systems are often considered
cognitive systems. However, these automated systems are not acceptable