Michal Kawulok, M.Sc. (2003), Ph.D. (2007), D.Sc. (2015), is an associate professor at the Silesian University of Technology (Gliwice, Poland) and a principal scientist at KP Labs (Gliwice, Poland). He has led numerous projects in both academia and industry, including those aimed at super-resolution reconstruction of satellite images, realized for European Space Agency. He has published over 100 papers in peer-reviewed journals and conference proceedings on pattern recognition, machine learning, and image processing. His research interests are concerned with super-resolution reconstruction, face and gesture recognition, linear and non-linear dimensionality reduction techniques, and support vector machines.
Keynote: Super-resolution for enhanced image understanding
Deep learning-powered super-resolution techniques allow for increasing spatial resolution of images and videos, generating visual content of extraordinarily high quality. Can they also lead to better performance of image analysis algorithms, if the latter are fed with the super-resolved images instead of the original ones? Can super-resolution serve as a pre-processing step to enhance the capabilities of automated image understanding, for example in the field of gesture recognition? In my talk, I will address these vital questions, focusing on the different approaches towards super-resolution, on the data used for training deep networks commonly employed for reconstruction, as well as on how to evaluate the super-resolution outcome.