Information societies technology

C7.3 UoS - University of Sussex (COGS)

Company profile - www.cogs.susx.ac.uk

The School of Cognitive and Computing Sciences (`COGS') at the University of Sussex grew out of a pioneering multi-disciplinary centre for research into intelligent systems and the mechanisms underlying them. It is one of the few centres in the UK where an interdisciplinary approach to the study of Computing and Cognitive Science is encouraged. Artificial Intelligence and Neuroscience are areas in which Sussex is exceptionally strong, a view supported by the results of the last three national Research Assessment Exercises in which the Computer Science and Artificial Intelligence (CSAI) group in COGS received the highest rating (5). The COGS research groups have made pioneering contributions to computer vision, genetic algorithms, artificial neural networks, simulation of adaptive behaviour, robotics and artificial life. A joint venture with COGS and the Centre for Neuroscience at BIOLS has led to the recent opening of the Sussex Centre for Computational Neuroscience and Robotics. The School includes research groups of international standing in the areas of Computer Vision, Neural Networks, and related Simulation of Adaptive Behaviour and Artificial Life who together share well equipped laboratory space.

Key persons

Hilary Buxton is Professor of Visual Intelligence, University of Sussex. She has a background in cognitive psychology, BSc (Bristol) and PhD (Cambridge). Her research has been developing computational theory for both human and machine vision systems, resulting in publication of a total of over 100 journal and conference papers. The work has involved parallel computer implementations of processing algorithms for both visualisation and interpretation of image sequences with much of the work being done in collaboration with UK and other European industrial partners. Previous project focused on biological motion understanding, medical visualisation, and algorithms for visual reasoning and control in advanced vision systems under ESPRIT project `Visual Inspection and Evaluation of Wide-area Scenes (VIEWS)’ and EPSRC project `Behavioural Analysis for Visual Surveillance'. She was principal investigator of EPSRC project `ISCANIT: Identity Recognition and Intentional Tracking' and EU project ‘PUPPET: Educational Puppet Theatre of Virtual Worlds’. She has new EU projects ‘ActIPret: Interpreting and Understanding Activities of Expert Operators’ as well as ECVision area leader. She has also work in conjunction with INRIA on the EU project ADVISOR to develop Bayesian learning techniques.
Christophe Ladroue started studying mathematics at Paris XII University, where he got

his BSc, with distinction. His dissertation was about the convergence of genetic algorithms, together with an application on the Iterated Prisoner Dilemma. After his military service, he went on for a Master in Artificial Intelligence and Pattern Recognition (DEA IARFA, Paris VII University). For his dissertation, he studied h fuzzy integrals and classifier combinations. From Jan 2000 onwards, he has been working within the European Project INTERPRET (with Tate). He is developing the pattern recognition tools used for the automatic classification (Support Vector Machines and

classical methods). In this work, emphasis is put on explanatory classifications when possible (in constrast to the 'black box approach') and on efficient and not misleading data display (representation of high dimensional objects in a 2D plan), for the end-user to make a quick decision with high confidence.

Selected References.

G. Ochoa, I. Harvey and H. Buxton. “On recombination and optimal mutation rates” Genetic and Evolutionary Computation Conference, GECCO, 1999.

G. Ochoa, I. Harvey and H. Buxton. “Error thresholds and their relation to optimal mutation rates” European Conference on Artificial Life, ECAL, 1999.
G. Ochoa, I. Harvey and H. Buxton. “Optimal mutation rates and selection pressure in genetic algorithms” Genetic and Evolutionary Computation Conference, GECCO, 2000.
A. R. Tate, J. Underwood, C. Ladroue et al."Visualisation of Multidimensional Data for Medical Decision Support" European Conference on Artificial Intelligence in Medicine, AIM, 2001.
A. R. Tate , M. Julia-Sape, C. Ladroue et al. "Automated Classification of Brain Tumours from 1H MR Spectra in INTERPRET, a Multi-Centre Collaboration", International Society of Magnetic Resonance Scientific Meeting, 2002.

C 7.4 University of Amsterdam (UVA)

Company profile - www.uva.nl

University of Amsterdam (UvA) in The Netherlands is one of the largest universities of the Netherlands, both in students as well as research. It will participate in the project through the Informatics Institute, of the Faculty of Sciences. The Institute has a well established tradition of European program participation on a broad range of topics.
Within the Informatics Institute, currently five research theme groups work on computer and network architecture, computational science, programming, distributed and federated information management, autonomous systems and sensory information analysis. On these topics, the institute covers the range

from fundamental research to application research in close collaboration with the industry.

ISIS group

The group on intelligent sensory information systems counts 25 members. Its primary goal is to create effective access to the content of digital images by bridging the semantic gap between pictorial data and the interpretation of the data. The gap is closed in from two sides: image data driven interpretation by computer vision from the one side, and knowledge driven from the other. The outcome of the research is algorithms and software of image analysis, experimenting on visual data, category learning from (very) large image databases, classification on the basis of known cases, performance analysis of algorithms and systems, and theoretical expansions. Where many items cannot be seen separated from their context, the scope of the analysis is expanded to full multi-media, but the emphasis remains on pictorial or video data. In the tradition of the institute, the research ranges from the theory of multimedia information analysis to innovative applications. Applications are in digital document structure analysis, video analysis, color image analysis and picture search engines. Application areas currently are in agricultural and industrial vision, document analysis and biological image processing.
The group on intelligent sensory information systems is involved in various industrial research projects, with Philips, Oce, Stork, Elsevier Science, and several others. Together with the Dutch center for Mathematics and Informatics (CWI), a grant of 4 MEuro with additional support from Elsevier, Oce, TNO, Datadistilleries and backed by research within the institutions aims at converting state of the art research in multi-media information analysis into demonstrators and half-products. The European IMAT project is developed with support from a variety of industries in Europe. We provide support in image analysis in the life sciences in a cooperation with institutes in Biology and Chemistry, as well as the National Brain Institute.