Prof. Nikos Komodakis
University of Crete, Greece
Time: 9 July PM
Discrete graphical models for medical image analysis: inference and learning methods
Abstract: Several problems in medical image analysis and computer vision can be formulated using the discrete graphical models framework. Such models are known to provide extreme flexibility and generality, and often lead to state-of-the-art solutions. The two main issues faced by researchers when using graphical models of this type are: (i) Learning: How to estimate the parameters of the model?; and (ii) Inference/optimization: How to find the best assignment for the variables of the model? The main aim of this tutorial will be to thoroughly discuss these two issues, starting from the basics and building up to the state of the art. Furthermore, it will discuss applications of these inference/learning methods to solving fundamental problems in medical imaging and computer vision.
Speaker Bio: Prof. Komodakis is currently serving as an adjunct professor at the Computer Science Department, University of Crete. He is also an affiliated researcher with INRIA-Saclay (the French research institute in informatics and control), working at the computer vision and medical imaging group. Prior to that he had also been a postdoctoral research fellow as well as an adjunct professor at the Applied Mathematics Department of Ecole Centrale de Paris (fellowship awarded by the Agence Nationale de la Recherche), which is one of the top three engineering schools in France (part of the elite of "Grande Ecoles"). In 2007, he was finalist for the prestigious ERCIM Cor Baayen Award for the most promising young researcher in computer science and applied mathematics.
Prof. Komodakis' research interests span various areas of multi-dimensional signal processing and analysis. In particular, he is interested in topics related to biomedical image analysis, computer vision, statistical pattern recognition, image processing, and machine learning. His work has appeared several times in the most reputed conferences and journals from the above fields. Together with his collaborators, he won best paper awards at IPMI 2007 and ISBI 2010. He also serves as a program committee member for a number of top international computer vision and pattern recognition conferences such as ICCV, ECCV and CVPR.