Variational methods are a flexible and powerful tool for the resolution of inverse problems in imaging.Contrary to a traditional approach which consists in reconstructing an image on a prixel grid, yielding a discrete problem, several recent methods such as the Beurling LASSO are directly formulated in a continuous domain.This generally yields solutions which are more faithful to the underlying physical problem, and the resulting discretizations may exploit the structures of the solutions (sparsity) and the continuous (or even differentiable) nature of the problem.
The aim of this workshop is to provide an overview of the latest advances on continuous methods for imaging.
Program The detailed program can be found here.
Slides The slides of some talks are available: click on 'Ordre du jour"/'Timetable' then click on the name of the speaker.
Registration is free but mandatory.
Kristian Bredies (University of Graz)
Blanche Buet (Université Paris-Saclay)
Lénaic Chizat (EPFL)
Clément Elvira (CentraleSupelec Rennes)
Maxime Ferreira Da Costa (CentraleSupelec)
Karl Hajjar (Université Paris-Saclay)
Clément Hardy (Ecole de Ponts Paris Tech)
Bastien Laville (Université Côte d'Azur)
Marta Lazzaretti (Universita degli studi di Genova/Université Côte d'Azur)
Romain Petit (Universita degli studi di Genova)
Clarice Poon (Warwick University)
Yann Traonmilin (CNRS/Université de Bordeaux)
Daniel Walter (Humboldt Universität zu Berlin)
Claire Boyer (Sorbonne Université)
Jean-Baptiste Courbot (Université de Haute Alsace)
Yohann De Castro (Institut Camille Jordan)
Vincent Duval (INRIA)
Emmanuel Soubies (CNRS, Université de Toulouse)
Funded by ANR CIPRESSI