Keynote Speakers
Title: Patch-based Inverse Problems in Interferometric Phase Imaging
Abstract: Interferometric phase imaging is a class of inverse problems aimed at the estimation of phase from sinusoidal and noisy observations. These degradation mechanisms (sinusoidal nonlinearity and noise) render interferometric phase imaging a quite challenging problem. In this talk I address two paradigmatic inverse problems of this class: a) interferometric denoising, which is the denoising of modulo-2pi phase images, and b) absolute phase estimation, which aims at recovering the original phase, including its 2pi-multiples. Interferometric denoising is tackled by reformulating the original estimation problem as a non-local patch-based inference problem in the complex domain.
Absolute phase estimation is tackled in two alternative ways: 1) via convex relaxation of the original problem; 2) via phase unwrapping, which is an integer optimization problem applied to the denoised interferometric images. The estimates produced by the addressed methods are characterized and their effectiveness illustrated in a series of experiments with simulated and real data.
Bio: Prof. Jose Bioucas-Dias received the EE, MSc, Ph.D., and Habilitation degrees in Electrical and Computer Engineering from Instituto Superior Tecnico (IST), Universidade Técnica de Lisboa (now Universidade de Lisboa), Portugal, in 1985, 1991, 1995, and 2007, respectively. Since 1995, he has been with the Department of Electrical and Computer Engineering, IST, where he is a Professor and teaches inverse problems in imaging and electric communications. He is also a Senior Researcher with the Pattern and Image Analysis group of the Instituto de Telecomunicações, which is a private non-profit research institution.
His research interests include inverse problems, signal and image processing, pattern recognition, optimization, and remote sensing. He has introduced scientific contributions in the areas of imaging inverse problems, statistical image processing, optimization, phase estimation, phase unwrapping, and in various imaging applications, such as hyperspectral and radar imaging. He is an IEEE Fellow and was included in Thomson Reuters' Highly Cited Researchers 2015 list and received the IEEE GRSS David Landgrebe Award for 2017.
Title: Geometric Representations of Graphs
Abstract: Intersection, contact or visibility representations of graphs are a natural way of graph and network visualization. At the same time the classes of graphs that arise in this way are popular and well studied for their interesting and elegant structural, as well as algorithmic properties. We will survey known results on these classes of graphs and discuss new as well as persisting open problems. The classes considered will include but not be restricted to interval, circle, circular arc, function and trapezoid graphs. The questions posed will concern recognition, complexity of optimization problems restricted to these classes, simultaneous representability and representability of planar graphs.
Bio: Prof. Jan Kratochvil has obtained his PhD at Charles University in Prague in 1987. There he is a Full Professor at the Department of Applied Mathematics since 2003. He has been a Fulbright Lecturer and a Visiting Associate Professor at the University of Oregon in 1990's, and a Visiting Professor at the University of Metz (2005) and Bordeaux University (2011). Since 2012 he is serving as the Dean of the Faculty of Mathematics and Physics at Charles University in Prague. Prof. Kratochvil's research interests lie in Discrete Mathematics and Theoretical Computer Science, namely structural and algorithmic properties of graph coloring and covering problems, visualization of graphs and domination in graphs. He is a steering committee member of WG - Graph Theoretical Concepts in Computer Science and GROW - Graph Classes, Optimization, and Width Parameters conference series.
Title: Representation learning with trainable COSFIRE filters
Abstract: In order to be effective, traditional pattern recognition methods typically require a careful manual design of features, involving considerable domain knowledge and expert efforts. The recent popularity of deep learning is largely due to the automatic configuration of effective early and intermediate representations of the data presented. The downside of deep learning is that it requires a huge number of training examples.
Trainable COSFIRE filters are an alternative to deep networks for the extraction of effective representations of data. COSFIRE stands for Combinations of Shifted Filter Responses. Their design was inspired by the function of certain shape selective neurons in areas V4 and TEO of visual cortex. A COSFIE filter is configured by the automatic analysis of a single pattern. The highly non-linear filter response is computed as a combination of the responses of simpler filters, such as Difference of (color) Gaussians or Gabor filters, taken at different positions of the concerned pattern. The identification of the parameters of the simpler filters that are needed and the positions at which their responses are taken is done automatically. An advantage of this approach is its ease of use as it requires no programming effort and little computation – the parameters of a filter are derived automatically from a single training pattern. Hence, a large number of such filters can be configured effortlessly and selected responses can be arranged in feature vectors that are fed into a traditional classifier.
This approach is illustrated by the automatic configuration of COSFIRE filters that respond to randomly selected parts of many handwritten digits. We configure automatically up to 5000 such filters and use their maximum responses to a given image of a handwritten digit to form a feature vector that is fed to a classifier. The COSFIRE approach is further illustrated by the detection and identification of traffic signs and of sounds of interest in audio signals.
The COSFIRE approach to representation learning and classification yields performance results that are comparable to the best results obtained with deep networks but at a much smaller computational effort. Notably, COSFIRE representations can be obtained using numbers of training examples that are many orders of magnitude smaller than those used by deep networks.
Bio: Prof. Petkov received his doctoral degree from Dresden University of Technology, Germany. Since 1991 he is a Professor of Computer Science (Chair of Intelligent Systems and Parallel Computing) at the University of Groningen, where he also served as scientific director of the Institute for Mathematics and Computer Science - now Johann Bernoulli Institute - from 1998 to 2009. He is also a member of the University Council and Chairman of the Science Faction since 2011. Petkov is an associate editor of several scientific journals and co-chair and co-organizer of international conferences. His research interests are in the field of development of pattern recognition and machine learning algorithms applied to various types of big data: image, video, audio, text, genetic, phenotype, medical, sensor, financial, web, and heterogeneous. He also develops methods for the generation of intelligent programs that are automatically configured using training examples of events and patterns of interest. Prof. Petkov served as Ph.D. thesis director (promotor) of 25 graduate students.