CISIM 2009 - Keynote Speakers
Gerald Schaefer
Department of Computer Science
Loughborough University
Loughborough, LE11 3TU,
United Kingdom
Title: Image database navigation: Visualisation, browsing and benchmarking
Abstract: Content-based image retrieval (CBIR) has been a very active research area over the last two decades with many interesting and promising approaches having been proposed that allow the retrieval of images based on features directly extracted from the images. However, since these features are typically low-level features, such as colour, texture or shape descriptors, it is hard to translate these into the higher-level semantic understanding of humans. Bridging this semantic gap still remains a challenge to be solved.
An interesting alternative to query-based approaches are image database navigation systems which allow the user to visually browse large image collections to arrive at images of interest in an effective, intuitive and efficient manner. The design of such browsing systems can be typically divided into two issues: visualisation of the image collection and the actual browsing of this visualisation.
Visualisation deals with the problem of presenting an image database in a way that the user is immediately able to grasp. The main difficulty here is of course how to deal with the large number of images that are present in the dataset. Various approaches to tackle this issue will be discussed. Mapping-based visualisations project thumbnail images into the typically 2-dimensional visualisation space in such a way that images which are visually similar are located close to each other in the visualisation. Clustered vsiualisations attempt to reduce the number of image that are required to be displayed at any one time by grouping similar together based on visual similarity. Graph-based visualisations utilise links between images to construct a graph where the nodes of the graph are the images and the edges the links between similar images.
Once an image database is displayed, the user should then be able to browse this collection. In order to be able to do this efficiently, various browsing operations are highlighted. These can be grouped into horizontal browsing operations which allow navigation within a single plane of visualised images, and vertical browsing which provides a means of navigating a hierarchical browsing structure.
Benchmarking image database navigation tools proves to be difficult due to the lack of a defined set of tasks and a standardised image dataset. A set of guidelines for benchmarking image browsing systems will be presented and discussed.
Biography: Gerald Schaefer gained his BSc. in Computing from the University of Derby and his PhD in Computer Vision from the University of East Anglia. He worked at the Colour & Imaging Institute, University of Derby (1997-1999), in the School of Information Systems, University of East Anglia (2000-2001), in the School of Computing and Informatics at Nottingham Trent University (2001-2006), and in the School of Engineering and Applied Science at Aston University (2006-2009) before joining the Department of Computer Science at Loughborough University in May 2009. His research interests are mainly in the areas of colour image analysis, content-based image retrieval, medical imaging and computational intelligence. He has published extensively in these areas with a total publication count of about 200. He is a member of the editorial board of several international journal, reviews for over 30 journals and serves on the programme committee of over 60 conferences. He is also the organiser of several international workshops and special sessions at conferences. His edited book on Computational Intelligence in Medical Imaging (CRC Press) has come out recently, and several further edited books are due to be published in 2009 and 2010.
Václav Snášel
Faculty of Electrical Engineering and Computer Science
VSB-Technical University of Ostrava,
Czech Republic
Title: Social Network Analysis
Abstract: Specific communication applications and devices such as email, instant messenger, blogs, discussion forum, and mobile telephony have led to an age of perpetual contact - social relations. Social Network Analysis (SNA) is the study of social relations among a set of actors. The key difference between network analysis and other approaches to social science is the focus on relationships between actors rather than the attributes of individual actors. A common framework for SNA is the mathematical approach of graph theory. This talk gives an overview of the basic SNA concepts.
Biography: Vaclav Snasel's research and development experience includes over 25 years in the Industry and Academia. He works in a multi-disciplinary environment involving artificial intelligence, multidimensional data indexing, conceptual lattice, information retrieval, semantic web, knowledge management, data compression, machine intelligence, neural network, web intelligence, data mining and applied to various real world problems. He has given more than 5 plenary lectures and conference tutorials in these areas. He has authored/co-authored several refereed journal/conference papers and book chapters. He has published more than 350 papers. He has supervised many Ph.D. students from Czech Republic, Jordan, Yemen, Slovakia, Ukraine and Vietnam. From 2001 he is a visiting scientist in the Institute of Computer Science, Academy of Sciences of the Czech Republic. From 2003 he is vice-dean for Research and Science at Faculty of Electrical Engineering and Computer Science, VSB-Technical University of Ostrava, Czech Republic. He is full professor since 2006. Before turning into a full time academic, he was working with industrial company where he was involved in different industrial research and development projects for nearly 8 years. He received Ph.D. degree in Algebra and Geometry from Masaryk University, Brno, Czech Republic and a Master of Science degree from Palacky University, Olomouc, Czech Republic. Besides, the Editor-in-Chief of two journals, he also serves the editorial board of some reputed International journals. He is actively involved in the International Conference on Computational Aspects of Social Networks (CASoN) ; Computer Information Systems and Industrial Management (CISIM); Evolutionary Techniques in Data Processing (ETID) series of International conferences. He is a Member of IEEE, ACM, AMS and SIAM.
Michal Wozniak
Faculty of Electronics
Wroclaw University of Technology
Wybrzeze Wyspianskiego 27,50-370 Wroclaw,
Poland
Title: "Classifier committees - how, why, and when they are working"
Abstract: Problem of pattern recognition is accompanying our whole life. Therefore methods of automatic classification is one of the main trend in Artificial Intelligence. The aim of such task is to classify a given object to one of predefined categories, on the basis of observing the features of the object. There is much current research into developing even more efficient and accurate recognition algorithms like neural networks, statistical and symbolic learning. Multiple classifier systems are currently the focus of intense research and in many review articles this trend has been mentioned as one of the most promising in the field of the pattern recognition. In the beginning in literature one could find only majority vote, but in later works more advanced methods of finding a common solution to the classifier group problem were proposed. There is a number of important issues while building the aforementioned multiple classifier systems like classifier selection for committee or choice of collective decision making method to name only a few. The proposed speech will present short review of the main methods of combined pattern recognition and their limits. The presented remarks will be illustrated by the results of experiments based on real classification problems.
Biography: Michal Wozniak is Professor of Computer Science in the Department of Systems and Computer Networks, Faculty of Electronics, Wroclaw University of Technology, Poland. He received an M.S. degree in Biomedical Engineering in 1992 from the Wroclaw University of Technology, Ph.D. and D.Sc. (habilitation) degrees in Computer Science in 1996 and 2007 respectively, from the same university. His research focuses on multiple classifier systems, machine learning, data and web mining, Bayes compound theory, distributed algorithms, computer and networks security and teleinformatics.
Prof. Wozniak has published over 120 papers, 2 books and edited 3 books Computer Recognition Systems (Springer). He is editor in chief of International Journal of Computer Networks and Communications and associate editor of several international journals including Pattern Analysis and Applications, Expert Systems, Computational Intelligence, Logic Journal of the IGPL, and International Journal of Communication Networks and Distributed Systems. He serves on program committees of numerous international conferences. His works have been transitioned into commercial applications. Prof. Wozniak has involved in many research projects related to machine learning, computer networks and telemedicine. Moreover, he has been consulting several commercial projects for the well known Polish companies and public administration. Prof. Wozniak is a member of IEEE (Computational Intelligence Society and Systems Man and Cybernetics Society) and IBS (International Biometric Society). See detail profile at: http://www.kssk.pwr.wroc.pl/pracownicy/michal.wozniak-en
Siti Mariyam Shamsuddin
Soft Computing Research Group,
Universiti Teknologi Malaysia,
Malaysia
Title: Moment-Based Discretization for Authorship Invarianceness
Abstract:
Biometric Technology has become an important research area now days. A
biometric system is fundamentally pattern recognition systems that work
by gaining biometric attributes from an individual, undergoing feature
extraction and evaluating the extracted feature set against the model
set in database. As such, the strong need to identify the right
authorship in fast, easy to use or for security experiences tremendous
growth, resulting in many applications being developed and
commercialized. The biometric data comprises of Handwriting, Signature,
Iris, Facial Features, Fingerprints and others. In this study, we
propose a new framework of Pattern Recognition with Discretization
Component being introduced to granularly mine the features of interest
for Authorship Invarianceness (as illustrated below).
In
our research, Authorship Invarianceness is obtained through proposed
integrated moment invariants and discretization scheme. The integrated
moment invariants are derived from Geometric Functions and United
Moment Invariants, while discretization scheme is done by dividing the
range of continuous attributes into disjoint regions (interval),and
labels are given to replace the actual data values. The proposed method
is categorized as supervised method since it needs class information to
perform discretization process. It globally process all integrated
invariants feature vector for all writers with dynamic characteristics
and search for the suitable set of cuts to represent the real data for
each writer. Later, the data are divided into the range of minimum to
maximum of each writer with equal size of interval or cuts. Number of
cuts is based on the number of feature vector for each word image, i.e,
eight feature vector values of proposed invariants are used to
represent a pattern image. This is aligned to the theory of geometric
moment function and to keep the original number of invariant vector in
moment invariant function that has been applied. The results show that
the proposed framework for writer identification with discretization
component has given significant impact on the performance of writers’
identification.
Biography:
Prof.
Dr. Siti Mariyam Shamsuddin received her Bachelor and Master degree in
Mathematics from New Jersey USA, and Phd in Pattern Recognition &
Artificial Intelligence from Universiti Putra Malaysia (UPM), MALAYSIA.
Currently, she is a Head, Soft Computing Research Group, K-Economy
Research Alliance, Universiti Teknologi Malaysia (UTM), Johor
MALAYSIA. She has published many journals and conference papers
theoretically and applications oriented. Her excellent works are
distinguished by a variety of award winnings from different exhibitions
nationally and internationally. Her research interests include Soft
Computing and its Application, Pattern Recognition & Forensic
Document Analysis, Geometric Modeling & Computer Graphics with Soft
Computing. She can be reached at mariyam@utm.my or
sitimariyams@gmail.com or website: http://gmm.fsksm.utm.my/~mariyam
Piotr Porwik, University of Silesia, Katowice, Poland
Title: Different methods of the fingerprints similarity measurement. A survey.
Abstract: Finger imprints are commonly used by police departments and many other civil areas of access protection. These methods still expand because
acquisition quality of devices (fingerprint scanners) is still growing – manufacturers improve resolution of such measuring instruments.
In addition, new algorithms based on modern image processing allow to extract the new type of fingerprint features.
Also new chemical methods allow to detect a fingerprint on materials, what has been very hard or impossible to achieve until now. It will be
more precisely explained and shown during the speech.
Bibliography: Piotr Porwik works in University of Silesia, Katowice, Poland as Head of Computer Systems Department. He received an M.S. degree in Computer
Science in 1979 from the University of Silesia, Ph.D. and D.Sc. degrees in Computer Science in 1985 and 2006 respectively, from
AGH-University of Science and Technology, Krakow, Poland. Prof. Porwik is reviewer of many International Journals and scientific societies. He has published over 100 papers. P. Porwik is editor in
chief of Journal of Medical Informatics and Technologies, and an associate editor of several other journals. Prof. Porwik’s main
scientific areas of interests are biometrics, and analysis of Boolean functions by means of spectral methods. More detailed info is
accessible from address http:/www.zsk.tech.us.edu.pl.