The task of image denoising is extensively studied for decades. A recent advance in this field results in dazzling results, so much so that some researchers believe that working on this problem leads to a dead avenue. Is this truly the case? What are the new trends in this field? Deep learning? New optimization methods? Can we leverage this impressive achievement and tackle other restoration problems, pushing these to new heights? All of these questions and more are the matter of this minisymposium.

Regularization by Denoising (RED): New and surprising uses of an old problem

Peyman Milanfar (Google Research)

Fast Algorithms for Regularization-by-Denoising

Phil Schniter (The Ohio State University)

Image Restoration by Iterative Denoising and Backward Projections

Raja Giryes (Tel Aviv University)

Divide and Conquer: Class-adapted Denoisers for Imaging Inverse Problems

Mário Figueiredo (Instituto de Telecomunicações and IST, University of Lisboa)

Learning to Mean-Shift in O(1) for Bayesian Image Restoration

Siavash Arjomand Bigdeli (EPFL)

Connect Maximum A Posteriori (MAP) Inference with Convolutional Neural Network for Image Restoration

Wangmeng Zuo (Harbin Institute of Technology)

On the Confluence of Deep Learning and Inverse Problems

Daniel Cremers (Technische Universität München)

Organizers:

Michael Elad (The Technion - Israel Institute of Technology)

Peyman Milanfar (Google Research)

Yaniv Romano (Technion - Israel Institute of Technology)

Keywords:

deep learning, image deblurring, image reconstruction, inverse problems, nonlinear optimization