Quantum Physics Inspired Image Processing: Image Restoration,
from Model-based to Learning Approaches
Abstract
Many signal or image processing tasks rely on linear model or some statistical assumptions on the observed signals/image or on the signal/image to be retrieved. In this talk, we will introduce a different way of image processing, based on quantum physics laws, referred to as quantum image processing (QuIP). We show that relating image samples to quantum features offers a great opportunity to efficiently address many tasks. Going a step further, QuIP is unfolded in a deep neural network via both quantum interactions and other quantum concepts, mainly the Hamiltonian operator. This leads to a versatile way to solve many inverse problems in image processing including denoising, deconvolution,… with interesting performances. Some applications are also discussed in the talk.
Short Biography
Denis Kouamé is a professor in medical imaging and signal processing at Paul Sabatier University of Toulouse, France. He currently leads the Signals and Images department at IRIT. His research interests are the general areas of Inversion in signal and image processing, computational imaging, medical imaging and namely medical ultrasound imaging, with special focus on image reconstruction and restoration. He has served as area chair or associate editor in several international conferences in signal, image processing or medical imaging, namely IEEE ICASSP, ICIP, ISBI,…. He is also regularly invited for talks in different universities inside and outside France. He was/is involved, as principal investigator or as member, in different European or French research projects (ANR, FUI, INSERM,…). His works resulted in creating different companies including two start-ups. He was an Associate Editor in Chief for IEEE Transactions on Ultrasonics Ferroelectrics and Frequency Control. He is a Senior Associate Editor for the IEEE Transactions on Image Processing and Associate Editor for the IEEE Transactions on Computational Imaging.