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
398 Synthesis and Analysis in Biometrics
Construction 1: I do not know, but something is not normal.
Construction 2: Possible alcohol intoxication.
Construction 3: An individual with a disability.
In this example, a 3-valued logic is used. Given m classes of
construction; an m-valued logic function can be applied. Generally,
psychologists must be consulted about the report formulation.
Note that this approach is a case of data mining, that is, a process of
nontrivial extraction of implicit and potentially useful information
[
34
]
.
15.5.5. Modularization
The modularity principle is implemented of all levels of the architecture. In
particular, the following modules, or Security Personnel Assistants (SPAs)
correspond to the specific biometric devices:
SPA for distance temperature measure,
SPA for distance blood pressure and pulse measure,
SPA for distance drug and alcohol intoxication estimation,
SPA for distance artificial accessory detection,
SPA for distance plastic surgery detection.
15.5.6. Estimations
The prototyping is also aimed at rendering the estimation techniques
introduced in Section 15.3. This can be accomplished by using a Bayesian
Belief Network (BBN) which describes the probability distribution of a set
of variables by specifying a set of conditional independence assumptions
together with a set of causal relationships among variables and their related
joint probabilities. The BBN is modelled by a direct acyclic graph, where
each node represents a variable to which a group of conditional probabilities
are attached. The BBN can be thought of as a probabilistic inference engine
that can answer queries, or What-If questions, about the variables in the
network. These queries include calculation of conditional probabilities, and
BBNs have the potential to make inferences under uncertain conditions.
Let X
1
,...,X
n
denote a set of n variables. The joint probability
distribution of variables in a direct acyclic graph is defined as p(X
1
=
x
1
,...,X
n
= x
n
)=
8
n
i=1
p(x
i
|x
∗
i
), where x
∗
i
is defined from the topology
of the graph (each node interacts with a bounded set of neighbors).
The following examples illustrate our approach to measurement of the
efficiency of analysis and processing data at various levels of a system.
April 2, 2007 14:42 World Scientific Review Volume - 9in x 6in Main˙WorldSc˙IPR˙SAB
Fundamentals of Biometric-Based Training System Design 399
Example 35. The effects of pre-screening can be estimated as the
difference of time, ∆T = T
2
− T
1
, and the difference of risks used to
make a decision, ∆I = I
2
− I
1
(Fig. 15.2(b)). The distance between the
pre-screening area and the authorization check can be used for obtaining
additional information (Fig. 15.17).
Effect
'I
I(t)
Time t
Pre
screening
Effect 'T
Screening
S
yste
m
A
S
yste
m
B
T
1
T
2
I
2
I
1
Prescreening
Gait biometric
I(t)
Time t
Authorized person
t
1
t
2
Screening
t
3
Unknown person
(a) (b)
Fig. 15.17. Screening T : effects of pre-screening can be estimated as the differences
∆T = T
2
− T
1
and ∆I = I
2
− I
1
(a), and, utilizing the topological and gait biometrics
in the training system (b).
Example 36. Figure 15.18 shows how the efficiency of questionnaire
techniques can be estimated using BBN.
Prescreening
information
I(t)
Time t
Questions Initiated by
biometric patterns
Confidence
interval for
decision making
Q5
Q2
Q4
Q3
Q1
Q7
Q6
Inefficient
questions or
answers
Fig. 15.18. Estimation of knowledge-based support using a function of insufficiency
I(t).
15.6. Discussion and Open Problems
We have developed a uniform and cost-effective platform of hardware and
software tools for PASS and for training users of PASS. We have introduced
April 2, 2007 14:42 World Scientific Review Volume - 9in x 6in Main˙WorldSc˙IPR˙SAB
400 Synthesis and Analysis in Biometrics
our experience and study in prototyping the main components of the
training system based on the PASS platform.
15.6.1. Novel Approaches
To achieve the study’s goals, a number of novel approaches were used
throughout design and prototyping, in particular:
Extension of the context of biometric data; we consider each biometric
device in a system as a smart sensor and an initiator for the
questionnaire.
Two-domain processing of biometric data; in addition to the signal domain
where biometric devices traditionally operate, we use the knowledge
domains, where biometric patterns are represented in semantic form.
Questionnaire techniques based on possible sources of information
(biometric data, personal data from databases, and the experience of
personnel)
Evaluation techniques are based on three main characteristics: degradation,
training, and evolution functions. These functions can be transformed
into each other and are well suited to the requirements of keeping
personnel skills at a high level of professionalism, due to their ability
to accept various strategies for improving the system.
Improving the reliability of biometric data. To alleviate the sensitivity
of biometrics to the environment, we have developed two approaches.
The first is based on a questionnaire technique that allows us to
“locate uncertainty” in processing data. The second approach is based
on correlation between biometric data in the signal and knowledge
domains.
Interpretability of biometric patterns insures human-friendly interface
when manipulating data, for example, derivations of the correlation
between biometrics.
In addition, we intensively used throughout our design the principle of
formidable resemblance, that is, the similarity of a set of algorithms based
on the scheme:
Data acquisition → Information extraction
→ Knowledge representation → Action
15.6.2. Open Problems
Several problems were identified through the prototyping of the system.
April 2, 2007 14:42 World Scientific Review Volume - 9in x 6in Main˙WorldSc˙IPR˙SAB
Fundamentals of Biometric-Based Training System Design 401
Problem 1. Extension of the pattern recognition problem for
various biometrics using a knowledge domain representation requires more
sophisticated tools. This is known as the integrated knowledge intensive
approach.
Example 37. One can observe the strong correlation between the facial
IR image and gait analysis patterns in semantic form:
Possible alcohol or drug intoxication.
Warning: level XX.
and
Gait: Possible alcohol or drug intoxication.
Warning: level 04.
The correlation is well identified in semantic form and can be used for
initiation of supporting questions as indicated in the protocol in subsection
15.4.3. This technique should be improved.
Problem 2. Alternative solutions to the questionnaire technique,
the basic component of decision making support technology, must be
investigated. In the prototyping system, logic equations for describing the
relationship between biometric patterns and sets of questions are currently
in use.
Example 38. The pattern “Temperature between 37 and 38”isassigned
to a value “1” of a four valued logic variable and corresponds to the semantic
construction “Do you need medical assistance?” However, the pattern
“Extremely high temperature” does not initiate a question, but rather an
action that corresponds to the early warning manifestation of the “Alarm”
scenario.
We have chosen this approach to knowledge representation and question
generation for several reasons. First of all, automatic minimization
of uncertainty and inconsistences is an approach that is still risky in
practice and costly to develop and implement
[
14,27
]
. Specifically, defined
inconsistencies were viewed through logical description and manifested to
the personnel to be resolved. Secondly, a questionnaire technique can be
adopted from well-studied polygraph systems
[
36
]
.
Problem 3. Synthetic biometric data generation. It is remarkable that
the next generation of security and training systems will be multi-biometric
and will use synthetic biometrics. However, the development of solutions
to inverse biometric problems (generating synthetic biometric data) and
April 2, 2007 14:42 World Scientific Review Volume - 9in x 6in Main˙WorldSc˙IPR˙SAB
402 Synthesis and Analysis in Biometrics
techniques for virtual environment design demand more effort. Our
prototyping is based on synthetic biometric data automatically generated
to “imitate” real data as described in
[
3,28,39,50,48
]
.
Example 39. Models can be generated from real acquired data and can
simulate age, accessories, and other attributes of the human face. This
environment, or personnel with virtual reality modeling, as well as it can
be used in decision making support.
Note that synthesis of biometric data brings up the classic problem of
computer graphics and computer vision: the generation of novel images
corresponds to an appropriate set of control parameters. In particular, in
[
3
]
, learning techniques were used given “pose” and “expression” parameters
for face synthesis.
We have not yet developed training scenarios using biometrics such
as synthetic iris
[
41,48,50
]
, ear topology
[
20,21
]
, and signature
[
32,37
]
.
Additional study is needed in the application of knowledge-based techniques
[
6
]
.
15.7. Conclusion
This study contributes to the development of a methodology for training
personnel in specific applications based on biometrics and security
screening. This research confirms that:
(a) Screening is a technique already adopted in social infrastructure
(for example, document-based checks, fingerprints, and mug-shot
acquisition),
(b) Screening can be efficiently combined with analytical, knowledge
intensive techniques,
(c) Screening is the basis for the implementation of the early warning
paradigm in authorization techniques, and
(d) Training systems can be developed based on the hardware and software
platform of the access control system.
Usage of various biometrics, including early detection and warning
devices in the system, must involve sociological and economic considerations.
The proposed decision making support system is considered to be a
unique area to combine and prototype various methodologies, concepts, and
techniques from many areas of the natural and social sciences, in particular:
image processing and pattern recognition, virtual environment design and
synthetic biometric data generation, distributed and multi-agent system
design, human-machine interaction, integrated knowledge-intensive system
design, communication, and psychology.
April 2, 2007 14:42 World Scientific Review Volume - 9in x 6in Main˙WorldSc˙IPR˙SAB
Fundamentals of Biometric-Based Training System Design 403
Acknowledgments
This Project is partially supported by the Natural Sciences and Engineering
Research Council of Canada (NSERC), the Canadian Foundation for
Innovations (CFI), the Government of the Province of Alberta, and ECSIS
— the European Center for Secure Information and Systems, Iasi, Romania.
A part of the project has been implemented as an initiative within the JPLs
Humanoid Robotics Laboratory.
The authors are grateful to many researchers in biomedical engineering,
knowledge-based decision making, security, and psychology, who provided
their valuable remarks and suggestions on this Project. Several commercial
companies provided their equipment and tools to prototype the described
system. The authors acknowledge the valuable suggestions and remarks of
Eur. Ing. Mr. Phil Phillips (Department of Information Security, Office
of National Statistics, UK), Dr. M. S. Nixon (School of Electronics and
Computer Science at the University of Southampton, UK), Dr. D. J. Hurley
(Analytical Engines Ltd., UK), Dr. P. Perner (Institute of Computer
Vision and Applied Computer Science, Leipzig, Germany), and Dr. R.
Elio (University of Alberta, Canada).
Bibliography
1. Anderson, D. N., Thompson, S. E., Wilhelm, C. E. and Wogman, N. A.
(2004). Integrating Intelligence for Border Security, J. of Homeland Security,
Jan., www.homelandsecurity.org
2. Bayarri, S., Fernandez, M. and Perez, M. (1996). Virtual reality for driving
simulation, Communication of the ACM, 39, 5, pp. 72–76.
3. Beymer, D., Shashua, A. and Poggio, T. (1993) Examples Based Image
Analysis and Synthesis, Technical Report 1431, Massachusets Institute of
Technology,20p.
4. Biometrics for Airport Access Control Guidance Package. Response to
Section 4011(a)(5), Version release Sept. 30, 2005, www.tsa.gov/interweb/
anetlibrary/Biometric
5. Burdea,G.C.andCoiffet,P.Virtual Reality Technology (2004). John Wiley
&Sons.
6. Coello Coello, C. A. (2006). Evolutionary algorithms: basic concepts and
applications in biometrics, this issue.
7. Crane, P. (1999). Implementing distributed mission training, Communication
of the ACM, 42, 9, pp. 90–94.
8. Davis, J. and Prosnitz, D. (2003). Technical and Policy Issues of
Counterterrorism - A Primer for Physicists, Physics Today, 56, 4, pp. 39–45,
April.
9. Dowdall, J., Pavlidis, I. and Bebis, G. (2003). Face detection in the near-IR
spectrum, Image and Vision Computing, 21, 7, pp. 565–578.
April 2, 2007 14:42 World Scientific Review Volume - 9in x 6in Main˙WorldSc˙IPR˙SAB
404 Synthesis and Analysis in Biometrics
10. Edwards, W. (1987). Decision making. In: G. Salvendy, Ed. Handbook of
human factors, Wiley, NY, pp. 1061–1104.
11. Elliott, S. J., Kukula, E. P. and Modi, S. K. (2006). Issues involving the
human biometric sensor interface, this issue
12. Eveland, C. K., Socolinsky, D. A. and Wolff, L. B. (2003). Tracking human
faces in infrared Vvideo, Image and Vision Computing, 21, pp. 579–590.
13. Flight simulation/simulators (1985), Society of Automotive Engineers,
Warrendale, PA.
14. Forsythe, C., Bernard, M. L., Goldsmith, T. E. Eds, Cognitive Systems:
Human Cognitive Models in Systems Design, (2006). Lawrence Erlbaum
Associates Publishers, New Jersey
15. Foster, J. P., Nixon, M. S. and Pr¨uugel-Bennett, A. (2003). Automatic gait
recognition using area-based metrics Pattern, Recognition Letters 24, pp.
2489-2497.
16. Fujimasa, I., Chinzei, T. and Saito, I. (2000). Converting far infrared image
information to other physiological data, IEEE Engineering in Medicine and
Biology Magazine, 10, 3, pp. 71–76.
17. Gaussorgues, G. (1994). Infrared Thermography, Microwave Technology
Series 5, Chapman and Hall, London.
18. Van Gemert, M. J. C., Jacques, S. L., Sterenborg, H. J. C. M. and Star,
W. M. (1989). Skin optics, IEEE Trans. Biomedical Engineering, 36, 12, pp.
1146–1154.
19. Herzog, G. and Reithiger, N. (2006). The SmartKom: Foundations of
multimodal dialogue systems. In: W. Wahlster, Ed., Springer, Heidelberg,
pp. 43–58.
20. Hurley, D. J., Nixon, S. M. and Carter, J. N. (2005). Force field feature
extraction for ear biometrics, Computer Vision and Image Understanding
98, pp. 491-512.
21. Hurley, D. J. (2006). Force field feature extraction for ear biometrics, this
issue
22. Jain, A., Pankanti, S. and Bolle, R. Eds. (1999). Biometrics: Personal
Identification in Networked Society, Kluwer, Dordrecht.
23. Jain, A. K., Ross, A. and Prabhakar, S. (2004). An introduction to biometric
recognition, IEEE Trans. on Circuits and Systems for Video Technology,
Special Issue on Image- and Video- Based Biometrics, 14,1.
24. Johannsen, G. and Rouse, W. B. (1983). Studies of planning behavior of
aicraft pilots in normal, abnormal, and emergency sumulations, IEEE Trans.
Systems, Man, and Cybernetics, 13, 3, pp. 267-278.
25. Keenan, P. (1998). Spatial decision support system for vehicle routing.
Decision Support Systems, 22, 1 pp. 65–71.
26. Knerr, B., Breaux, R., Goldberg, S., Thrurman, R. (2002). National defence.
In: K. Stanney, Ed. The Handbook of Virtual Environments Technology,
Erlbaum, Mahwah, NJ, pp. 857–872.
27. Lapeer, R. J., Chios, P., Linney, A. D., Alusi, G. and Wright A. (2001). HCI:
The next step towards optimization of computer-assisted surgical planning,
intervention and training (CASPIT), In: Human Computer Interaction:
April 2, 2007 14:42 World Scientific Review Volume - 9in x 6in Main˙WorldSc˙IPR˙SAB
Fundamentals of Biometric-Based Training System Design 405
Issues and Challenges, Chen Q. Ed., Idea Group Publishing, Hershey, PA,
pp. 232–246.
28. Ma, Y., Schuckers, M. and Cukic, B. (2005). Guidelines for Appropriate
Use of Simulated Data for Bio-Authentication Research, Proc. 4th IEEE
Workshop Automatic Identification Advanced Technologies, Buffalo, NY, pp.
251–256.
29. Mather, G. and Murdoch, L. (1994). Gender discrimination in biological
motion displays based on dynamic cues. Proc. Royal Society, 258, pp. 273–
279.
30. Method to Testing Objects for Random Actions Thereon and Digital
Simulator-Computer System for Effecting Same, Patent U.S.A. 4205383,
Bakanovich,E.A.,Shmerko,V.P.atal.
31. Nixon, M. S., Carter, J. N. Cunado, D., Huang, P. S. and Stevenage, S. V.
(1999). Automatic gait recognition. In: A. Jain, R. Bolle, and S. Pankanti,
Eds, Biometrics: Personal identification in a networked society,Kluwer,
Dordrecht, pp. 231–250.
32. Parker, J. R. (2006). Composite systems for handwritten signature
recognition, this issue
33. Pavlidis, I. and Levine, J. (2002). Thermal image analysis for polygraph
testing, IEEE Engineering in Medicine and Biology Magazine, 21,6,pp.
56–64.
34. Perner, P. (2005). Case-based reasoning for image analysis and interpretation,
In: C.ChenandP.S.P.Wang,Eds.Handbook on Pattern Recognition and
Computer Vision, 3rd Edition, World Scientific Publisher, pp. 95–114.
35. Pleban, M., Niewiadomski, H., Buclak, P. Sapiecha, P. Yanushkevich, S.
and Shmerko, V. (2002). Argument Reduction Algorithms for Multi-Valued
Relations, Proc. Int. Conf. Artificial Intelligence and Soft Computing, Banff,
Canada, pp. 609–614.
36. The Polygraph and Lie Detection, (2003). The National Academies Press,
Washington, DC, 2003
37. Popel, D. V. (2006). Signature analysis, verification and synthesis in pervasive
environments, this issue
38. Prokoski, F. J. and Riedel, R. B. (1999). Infrared identification of faces
and body parts, In: A. Jain, R. Bolle, and S. Pankanti, Eds., Biometrics:
Personal Identification in Networked Society, Kluwer, Dordrecht, pp. 191–
212.
39. Ratha, N., Connell, J., and Bolle, R. (2001). Enhancing security and privacy
in biometrics-based authentication systems, IBM Systems J., 40, 3, pp. 614–
634.
40. Richards, A. (2001). Alien Vision: Exploring the Electromagnetic Spectrum
with Imagin Technology, SPIE.
41. Samavati, F. F., Bartels, R. H. and Olsen, L. (2006). Local B-spline
Mmultiresolution with example in iris synthesis and volumetric rendering,
this issue.
42. Shmerko, V, Perkowski, M., Rogers, W., Dueck, G., and Yanushkevich,
S. (1998). Bio-Technologies in Computing: The Promises and the Reality,
April 2, 2007 14:42 World Scientific Review Volume - 9in x 6in Main˙WorldSc˙IPR˙SAB
406 Synthesis and Analysis in Biometrics
Proc. Int. Conf. on Computational Intelligence & Multimedia Applications,
Australia, pp. 396–409.
43. Srivastava, A. and Liu, X. (2003). Statistical hypothesis pruning for
identifying faces from infrared images, Image and Vision Computing, 21,
pp. 651–661.
44. Sugimoto, Y., Yoshitomi, Y. and Tomita, S. (2000). A Method for detecting
transitions of emotional states using a thermal facial image based on a
synthesis of facial expressions, Robotics and Autonomous Systems, 31, pp.
147–160.
45. Total Information Awareness DAPRA’s Research Program, (2003).
Information and Security, 10, pp. 105–109.
46. Tsumura, N. Ojima, N. Sato, K., Shiraishi, M. Shimizu, H., Nabeshima,
H., Akazaki, S., Hori K. and Miyake, Y. (2003). Image-based skin color and
texture analysis/synthesis by extracting hemoglobin and melanin information
in the skin, ACM Trans. on Graphics, 22, 3, pp. 770–779.
47. Wang, W., Zeng, Y., Ma, D., Jin, Z., Wu, H., Yuan, C., Yuan, Y., Lin, Z.,
Wang, C. and Qi, H. (2004). Clinical study on using thermal texture maps
in SARS diagnosis, Proc. 26th Annual Int. Conf. of the IEEE EMBS,San
Francisco, CA, pp. 5258–5264.
48. Yanushkevich, S. N., Stoica, A., Shmerko, V. P. and Popel, D. V. (2005).
Biometric Inverse Problems. CRC Press/Taylor & Francis Group, Boca
Raton, FL.
49. Yanushkevich, S. N., Miller, D. M., Shmerko, V. P. and Stankovi´c, R. S.
(2006). Decision Diagram Techniques for Micro- and Nanoelectronic Design,
Handbook, Taylor & Francis / CRC Press, Boca Raton, FL.
50. Yanushkevich, S. N. (2006). Synthetic biometrics: A survey, Proc. World
Congress on Computational Intelligence, Vancouver, Canada.
51. Yanushkevich, S. N., Stoica, A. and Shmerko, V. P. (2006). Experience
of Design and Prototyping of a Multi-Biometric, Physical Access, Early
Warning Security Systems Based on Screening Discipline of Service
(PASS) and Training System T-PASS, Proc. 32nd Annual IEEE Industrial
Electronics Society Conf., Paris.
52. Yanushkevich, S. N. and Stoica, A. (2006). Supervisors, Design and
Prototyping of a Multi-Biometric Physical Access Screening Early Warning
Security Systems (PASS) and Training System T-PASS, Technical Report
ECE2006.BT.R04, Biometric Technology Laboratory, University of Calgary
(Canada) and Humanoid Robotics Laboratory, California Institute of
Technology, NASA Jet Propulsion Laboratory, California Institute of
Technology Pasadena, CA (USA).
53. Zhang, D. (2000). Automated Biometrics: Technologies and Systems,Kluwer,
Dordrecht.
54. Zhang, D. (2004). Palmprint Authentication, Kluwer, Dordrecht.
55. Zellner, A. (1988). Optimal information processing and Bayes’ theorem,
American Statistician, 42, 4, pp. 278–284.
April 2, 2007 14:42 World Scientific Review Volume - 9in x 6in Main˙WorldSc˙IPR˙SAB
About the Authors and Contributors
Ahmed Awad E. Ahmed is a research associate
and a Ph.D. candidate at the Electrical and Computer
Engineering Department, University of Victoria. His Ph.D.
thesis presents a number of new trends in security
monitoring through human computer interaction devices.
He is a member of the Security and Object Technology
(ISOT) Research Group and the principal investigator
of Biotracker, an intrusion detection system based on
biometrics. Mr. Ahmed worked as a technology specialist and a quality assurance
consultant in a number of leading software firms. He obtained his B.Sc. and M.Sc
from Ain Shams University, Electrical and Computer Department on 1992, and
1997 respectively.
Richard Harold Bartels born in Ann Arbor, Michigan,
in 1939, Bartels Bartels earned his Bachelor and Master
degrees from the University of Michigan in 1961 and 1963,
respectively. He earned his PhD from Stanford University
in Computer Science with a specialization in numerical
analysis in 1968. He worked for the Computer Science
Department and the Center for Numerical Analysis at
the University of Texas and Austin from 1968 through
1974, then for the Mathematical Sciences Department of The Johns Hopkins
University in Baltimore, Maryland, until 1979, and finally for the Computer
Science Department of the University of Waterloo in Ontario, Canada, until his
retirement in 1999. He has authored papers in numerical linear algebra, numerical
optimization, and computer-aided geometric modeling. He is the co-author of a
book on the use of splines in computer graphics.
407