Engineering Solutions for Automatic
Identification of Objects and Persons
Ph.D. Thesis
László Kutor
Advisor
:
Dr. György Kozmann
University of Pannonia
Faculty of Information Technology
Information Science and Technology PhD School
2007
1. Automatic identification of objects and persons:
Importance, research background and objectives
The effective and reliable identification of objects and people - a natural and necessary process that we apply in varying degrees to virtually everything we interact with in life - is a critical prerequisite for today’s increasingly ubiquitous IT systems. The effective operation of today’s complex and interconnected electronic elements and IT systems is predicated upon their ability to reliably – and in most cases automatically – identify and authenticate objects and individuals. Authentication is, for example, required to control access to system resources, to authorize system use, and to later account for system usage – all interrelated functions.
One clear sign of the importance of the automatic identification of objects and people is the intense level of research being conducted in this field, for which there is a wide and heterogeneous body of literature. The seemingly unlimited variety of applications, diversity of differentiating characteristics that serve as the basis for authentication, and the wide range of methods used to record and analyze these characteristics makes each system appear to be disparate and unique.
The first research objective of this dissertation is to examine the critical
factors in automatic authentication and to define a systematized approach. This new approach divides the characteristics of the objects, persons or systems to be identified and authenticated into two main groups:
(1) Identification properties based on innate, internal and inseparable characteristics,
(2)Assigned identification codes
The use of identification properties based on innate characteristics is preferable to the use of assigned codes for several different reasons, including tradition, ease of use, and efficiency. However, they also typically require an intense level of computations and a system that is capable of doing this in real time. For this reasons, most real-life applications tend to rely on the assigned code method. Using this typology, the dissertation presents the most promising and spreading identification and authentication methods, including digital signatures, biometric passports, applications
based on Radio Frequency IDentification (RFID), and biometric
authentication of persons.
The second research objective of this dissertation is to present two
pioneering applications using original automatic identification and authentication methods developed during the research process. The first is an integrated hardware/software system designed to accurately capture and identify complex images of medical operations based on local information. The second application automatically records, analyzes and authenticates in real time an individual’s signature. The original authentication process incorporates an analysis of the graphical structures (size and shape) of the signature elements and augments this, using biometric methods, with a precise analysis of the dynamic characteristics such as pressure, pen angle, and pen movement in the air and on the writing surface.
2. The research approach and phases
The original methods of identification and authentication were developed through a lengthy process of applied informatics research. The overall aim of the research was to develop the optimal solution to real-life problems. Following a thorough review of the relevant theory an existing literature, the research sought to answer the question:
“How can optimal solutions to real-life authentication challenges be
developed incorporating the most effective and reliable authentication methodology and state of the art technology such as RFID and biometrics?”
Consistent with best practices of engineering science, the research methodology consisted of the following steps:
• Definition of the technical problem to be solved and identification of the critical factors,
• Review of relevant literature and research aimed at solving the same problem,
• System design, with justification of the decisions and trade-offs made during this process,
• Development of working prototype to test the hypothesis,
The two primary areas of research focus were:
(1) The challenges involved in capturing the data needed for identification and authentication, and selecting the optimal distinguishing characteristics,
(2) The optimal authentication methodology based on the selected distinguishing characteristics. The main factors driving the development of authentication methods were effectiveness and feasibility.
The review of the literature on capturing and analyzing the distinguishing characteristics typically revealed that the methods and technology available were very much in the initial stages of development or constituted proprietary information. Both the recording instruments used to capture the data and the analytical methods typically required substantial additional development and testing.
In the course of the research process, I developed numerous authentication databases. For example, a database was developed for fingerprint identification that included the results from using two different scanners to record the prints of all ten fingers of several hundred individuals.1 Another database was created for the analysis of images of the iris.2 A third database was created for signature authentication that contained 30 signature samples from each of the several hundred test individuals.3 In addition, several thousand images were recorded during a series of autopsy procedures.4
During research into RFID technology, I also tested several different RFID tags and scanners to determine the optimal solution.5
Finally, I conducted additional research in authentication methods as part of six large research and development projects and served as the research advisor to 29 informatics college students who prepared undergraduate theses and conducted additional research in the field of identification and authentication.
1 Completed on commission from and with the collaboration of Dermotrade
2 With the permission and using equipment provided by the Opthmalogy Clinic of the Semmelweis Medical University.
3 In cooperation with the Institute of Graphology.
4 In cooperation with the Institute of Anatomy of the Semmelweis Medical University 5 In cooperation with Vonalkód Rendszerház
3. Original research results
The original research findings presented in detail in the dissertation can be summarized in the following:
1. I developed a new integrated system for capturing images during medical operations and simultaneously reconstructing and
combining these images based on the precise identification of their exact position.
1.1 I designed a methodology and prototype technology that enables these images to be captured and reproduced from any angle within a 140 degree circumference.
1.2 I designed and created a highly reliable system for measuring the optical angle of the images, which enabled the images and the angle from which they were taken to be identified with a 0.1 degree of precision.
2. I developed a new method for identifying persons based on real-time analysis of the dynamic signature characteristics.
2.1 I developed the signature digitalizing process that made it possible to capture the dynamic characteristics in real time.
2.2 I developed the process for calibrating the instrument used to digitalize the signatures, which was originally designed for digitalization of drawings.
2.3 I determined the set of static and dynamic characteristics to be captured and analyzed.
2.4 I designed the method used to capture the signature, and the
corresponding reference database used in the process of authentication. 2.5 I designed the methods used to calculate, compare, and present the dynamic signature characteristics.
2.6 I designed the algorithm used to authenticate the signature based on the analysis of the set of characteristics.
4. Practical applications
4.1 As part of this research, I completed a working prototype of an
integrated hardware/software system that can identify and reproduce the exact position of anatomical images. The prototype has been in use at
the Institute of Anatomy, Pathology and Development of the Semmelweis Medical University since November 2006. This system has been used to record images of the human anatomy taken during autopsies. The images have then been compiled into initial modules of an interactive electronic atlas of human anatomy known as 4D Anatomy.
(www.4danatomy.com/home.html). The image identification methodology I developed made it possible to clearly identify the precise area of operation and the position of the anatomical structures and to view them from virtually any angle in an interactive manner.
The modules of the anatomical atlas completed to date were presented at an international medical technology exhibit in Las Vegas in May 2007 and were well received. As a result, several universities in the United States as well as the Karolinska Institute in Stockholm, Sweden, expressed their interest in using the software. This software system, which is based on the image recording and identification technology that I developed, may well play a significant role in improving the training of doctors and as a means to document operating procedures in the future.
4.2 Automatic and reliable signature authentication is a fundamental
requirement in the financial and retail industries, the lack of which causes billions of dollars of damage each year. As a first step, many banks and stores have introduced electronic devices that record client signatures. In the United States, for example, numerous retail stores require that check or credit card payments be made using electronic signatures. However, in the absence of appropriate technology, these signatures are typically still
authenticated through visual means by the cashiers. To address this need, a prototype of a signature authorization system was developed as part of a project supported by an IKTA grant. One major component of the signature authentication method used by this prototype is based on my research in the dynamic characteristics of signatures.
5. List of related publications
Refereed international journal articles
1. Balogh, A., Preul, M.C., Kutor, L., Schornak, M., Hickman, M., Deshmukh, P., Spectzler, R., “Multilayer Image Grid Reconstruction Technology: Four-Dimensional Interctive Image Reconstruction of Microsurgical Neuroanatomic Dissections,” Operative Neurosurgery 2006, January Vol. 58. 157-165. (IF=2.338;
2005)
2. Kozlovszky, M., Berceli, T., Kutor, L., “Analysis of SCTP and TCP based communication in high-speed clusters,” Nuclear Instruments and Methods in Physics Research A 559 (2006) 85-89. (IF=1.166; 2005)
3. Cserjés,A., Kutor, L., Izsó, L. “Identifying person-specific dynamic patterns in handwritten signatures by Bayesian neural networks,” International of Psychology,
(2004).39 (5-6): 70-72 Suppl. S OCT-DEC 2004.
Presentations at international conferences
4. Kutor, L., Dorkó, Gy., Ladányi, Z., “Íráskép vagy dinamika? Automatikus személyazonosítás a kézírás alaki és dinamikus jellemzői alapján.” (Image or dynamics? Automatic authentication of signatures based on graphic structures and dynamic characteristics)
Nemzetközi grafológiai kongresszus. 1999. Proc.:83-86
5. Kutor, L., Ladányi, Z., “Problems and Solutions in Dynamic Signature Authentication,” International Conference on
Computational Cybernetics, Hungary, 2003. ISBN 963 15418 3
6. Cserjés, Á., Kutor, L., “Identifying person-specific dynamic patterns in handwritten signatures by Bayesian neural networks,” 28th
International Congress of Psychology, Beijing, China, 2004.
7. Kutor,L., “Near Field Communication and Potential Implications,”
International conference on privacy issues in electronic identification
systems: Technical and legal questions on RFID. Centre for Infocommunications
Law (CIL), Hungarian Academy of Sciences, Institute of Legal Sciences, Hungary, 2006.
8. Kutor, L., “Store Logistics and Payment with Near Field
Communication,” First application conference on Near Field Communication,
9. Benyó, B., Vilmos, A., Kovacs, K., Kutor, L., “NFC Applications and Business
Model of the Ecosystem,” 16th IST Mobile & Wireless Communications Summit, Hungary 2007.
10. Benyó, B., Vilmos, A., Kovács, K., Kutor, L., “The Design of NFC Based Applications,” INES 2007, 11th International Conference on Intelligent Engineering Systems, 2007. Proc. pp. 277-280.
11. Kutor, L., Domoz, Zs., Nagy, G., Peller, Cs., Véső, T., Object Identification and Local Information Services Using Near Field
Communication RSEAS 2007, Regional Conference on Embedded and Ambient Systems, 22-24 November 2007. Budapest
Participation in related scientific research projects
1. Advisor to 29 informatics students at the Applied Informatics Institute of the Budapest Technical College, Neuman Faculty of Informatics, who conducted research in automatic authentication as the basis for their undergraduate theses or research projects.
2. Institutional research leader in the following additional research projects: • IKTA 88/2001 Design of original signature authentication system. 2001-2004;
Coordinator: Hungarian Institute of Graphology
• GVOP 3.1.1 Development of cell phone with Text-to-Speech functions. 2005-2007; Coordinator: MIT Systems
• GVOP 3.3.3 Market introduction of RFID technology through the establishment of a new inventory system and a pilot technical laboratory.
2005-present; Coordinator: Vonalkód Rendszerház
• 4D system for recording and reproducing anatomical images. 2004-present; Coordinator: 4D Anatomy
• GVOP GVOP 3.3.3. Development of monitoring information systems using new generation technology for use in facility management. 2006 – present; Coordinator:
Regens Rt.
• FP 6 „StoLPaN” PROJECT NO. : IST – 033591” Turning NFC enabled mobile handsets into multifunction terminals with bi-directional interaction between the NFC chip and wireless communication channels.” 2006 – present; Coordinator:
