Eric Pauwels

Scientific Staff Member, Coordinator Partnerships & Transfer, Group leader

Full Name Dr. E.J.E.M. Pauwels
Function(s) Scientific Staff Member, Coordinator Partnerships & Transfer, Group leader
Telephone +31 20 592 4225
Room M364
Department(s) Intelligent and Autonomous Systems

Research

My current research focuses mainly on two topics. On the one hand, I am working on content-basedimage retrieval where we are developing mathematically principled methodologies for characterising the geometrical and statistical properties of image models that are directly tied to perceptually relevant visual content.This builds on my earlier interests as it necessitates a multi-disciplinary approach which draws onfields as diverse as PDEs, information theory, data mining and machine learning. As for the latter, we are particularly interested in the application of statistical learning to data-driven or adaptive multimedia processing. The expertise thus developed is being applied toimage mining for databases in art and bio-diversity.

The second main theme concerns ubiquitous computing and pervasive sensing in heterogeneous sensor networks. There are many applications in which bottom-up processing for a single modality (such asvision) proves to be far too brittle. Robustness can be improved by injecting prior or high-level knowledge (e.g. certain geometric configurations are more likely than others) and/or by incorporating information gleaned from other sensors (e.g. combining information from voice and face recognition modules). I am particularly interested in the characteristics of the underlying computational framework that will give rise to robust and self-organising behaviour of such sensor networks. The aim is to develop methodologies that will allow the network to learn from multipledatastreams, adapt to changes in the environment and self-(re)organise when sensors are added or removed from the network. Again, this gives rise to interesting statistical problems, particularly in Bayesian reasoning, classification and active learning.

Recent Publications

  • B.J. Liefers, F.N. Claessen, E.J. Pauwels, P.A.N. Bosman, J.A. La Poutré, Market Garden: a Simulation Environment for Research and User Experience in Smart Grids, Springer, 351-354, 2014
  • C. Do, E.J. Pauwels, Using MathML to Represent Units of Measurement for Improved Ontology Alignment, Springer, 1-325, 2013
  • C. Do, E.J. Pauwels, Harnessing Mathematics for Improved Ontology Alignment, Springer, 1-318, 2013
  • J. Han, E.J. Pauwels, P.M. de Zeeuw, Fast Saliency-aware Multi-modality Image Fusion. , Elsevier, 70-80, 2013
  • J. Han, E.J. Pauwels, P.M. de Zeeuw, Visible and Infrared image registration in man-made environments employing hybrid visual features. , north-holland, 42-51, 2013