Visit of professor Gabriele Steidl

During the visit of Gabriele Steidl, new solutions for restoration problems involving images corrupted with Poisson noise have been devised. The idea is to employ an Anscombe based approximation which allows the variational problem to be formulated as a constrained convex optimization problem. The constraint is addressed through a novel epigraphical techniques and recent primal-dual schemes. In the continuation of this work, Audrey Repetti (PhD student of J.-C. Pesquet) has spent one month in the team of Gabriele Steidl where she has investigated proximal properties of so-called quotient functions which are of interest in optimization problems for database management.The corresponding work will be published in 2015 in the J. of Comp. and Appl. Math. : « Proximal operator of quotient functions with application to a feasibility problem in query optimization » by G. Moerkotte, M. Montag, A. Repetti and G. Steidl.