Chair
Ana Fred
IT - Instituto de Telecomunicações
IST - Technical University of Lisbon
Portugal
e-Mail
Aim of the Workshop
Huge amounts of information in digital format are currently available; for example, the world-wide-web is an immense source of text, audio, images and video. Organizing and understanding this multimodal information is a difficult problem, which requires advanced techniques from pattern recognition and related fields such as machine learning, computer vision and human language technology. An increasingly important aspect of this problem is human-computer interaction, especially multi-modal interactive systems in which humans and machines seamlessly cooperate to efficiently solve a given task. On the other hand, accessibility issues also arise in this problem, with personal identification techniques playing a key role in human-computer interfaces for security and restricted-access systems. The aim of this workshop is to bring together researchers, practitioners and potential users with interests in the multidisciplinary topics listed below.
Topics of Interest
Paper submissions should be related to the following areas: pattern recognition (PR), machine learning, computer vision and human language technology. Topics of interest include, but are not limited to:
- Foundations of PR: feature-based and similarity-based PR
- Novel PR algorithms
- Distributed data analysis
- Scalable PR
- Unsupervised learning and clustering
- Classification and browsing of multimodal data
- Indexing, structuring and summarization
- Data reduction, selection and transformation
- Data visualization
- Privacy preserving data analysis
- Text, audio, image and video processing
- Web mining
- Web, social networks and communities
- Human-computer interaction modelling
- Discourse and dialogue
- Machine translation
- Multimodal fusion
- Biometrics and personal authentication systems
Keynote Speakers
Robert P. W. Duin, TU Delft, The Netherlands
Title: Pattern Recognition as a Human Centered non-Euclidean Problem
Abstract
Regularities in the world are human defined. Patterns in the observed phenomena are there because we define and recognize them as such. Automatic pattern recognition has thereby to bridge the gap between human judgment and measurements made by artificial sensors. A good, well performing pattern recognition system is adapted to what we experience as similar and as dissimilar. A natural way to design such a system is thereby based on the pair wise comparison of the objects for which patterns should be recognized. This results in the dissimilarity representation for pattern recognition. An analysis of such representations optimised for performance shows that they tend to be non-Euclidean. The Euclidean vector spaces, traditionally used in pattern recognition and machine learning are for these cases suboptimal. The causes and consequences of the use of non-Euclidean representations will be discussed.
Short Biography
Robert P.W. Duin received in 1978 the Ph.D. degree in applied physics from Delft University of Technology, Delft, The Netherlands, for a thesis on statistical pattern recognition. He is currently an Associate Professor in the Faculty of Electrical Engineering, Mathematics and Computer Science of the same university.
During 1980-1990, he studied and developed hardware architectures and software configurations for interactive image analysis. After that he became involved with pattern recognition by neural networks. His current research interests are in the design, evaluation, and application of algorithms that learn from examples, which includes neural network classifiers, support vector machines, classifier combining strategies, and one-class classifiers.
Especially complexity issues and the learning behavior of trainable systems receive much interest. Recently, he started to investigate alternative object representations for classification and he thereby became interested in dissimilarity-based pattern recognition, trainable similarities, and the handling of non-Euclidean data.
Dr. Duin is an associated editor of Pattern Recognition Letters and a past-associate editor of the IEEE Transactions on Pattern Analysis and Machine Intelligence. He is a Fellow of the International Association for Pattern Recognition (IAPR). In August 2006 he was the recipient of the Pierre Devijver Award for his contributions to statistical pattern recognition.
Workshop Program Committee
Selim Aksoy, Bilkent University, Turkey
Alberto Del Bimbo, University of Florence, Italy
José Bioucas-Dias, Instituto Superior Técnico, Lisboa, Portugal, Portugal
Julian Fierrez, Universidad Autonoma de Madrid, Spain
Mario Figueiredo, Instituto Superior Técnico, Portugal
Jose M. Iñesta, Universidad De Alicante, Spain
Alfons Juan, Universitat Politècnica de València, Spain
Jean-Marc Ogier, Laboratoire Informatique, Image, Interaction (l3i), France
Marcello Pelillo, University of Venice, Italy
Petra Perner, Institute of Computer Vision and Applied Computer Sciences, Germany
Arun Ross, West Virginia University, U.S.A.
Lilian Tang, University of Surrey, U.K.
Eric Trupin, University of Rouen, France
Aditya Vailaya, Retrevo, Inc., U.S.A.
Enrique Vidal, Universidad Politécnica De Valencia, Spain
(list not yet complete)
Publications
All accepted papers will be published in the workshop proceedings book, under an ISBN reference, and on CD-ROM support. The proceedings will be available at the time of the workshop.
