Image denoising has reached impressive heights in performance and quality -- almost as good as it can ever get. This talk is about the many other things one can do with a good denoiser besides using it for its intended purpose. Of particular interest is how to use denoisers in the regularization of any inverse problem. We propose an explicit image-adaptive regularization functional that makes the overall objective functional clear and well-defined. Remarkably, the resulting regularizer is convex. With complete flexibility to choose the iterative optimization procedure for minimizing this functional, RED is capable of incorporating any image denoising algorithm as a regularizer, treat general inverse problems very effectively, and is guaranteed to converge to the globally optimal result. I will show examples of applications, including tone-mapping, deblurring, and super-resolution.

This presentation is part of Minisymposium “MS52 - A Denoiser Can Do Much More Than Just... Denoising (2 parts)”
organized by: Yaniv Romano (Technion - Israel Institute of Technology) , Peyman Milanfar (Google Research) , Michael Elad (The Technion - Israel Institute of Technology) .

Authors:

Michael Elad (The Technion - Israel Institute of Technology)

Peyman Milanfar (Google Research)

Yaniv Romano (Technion - Israel Institute of Technology)

Keywords:

deep learning, image reconstruction, inverse problems