Shearlab.jl (https://github.com/arsenal9971/Shearlab.jl) is a Julia Library with a toolbox for multidimensional data processing using the shearlet system as representation system, which is known to provide optimally sparse approximation of functions governed by anisotropic features. This talk will be focused on the use of the library for solving ill-posed inverse problems in imaging by sparse regularization applied to problem classes such as image denoising and image inpainting. Finally, we will discuss how to optimize the implementation using gpu accelerators.

This presentation is part of Minisymposium “MS76 - Solving Inverse Problems in minutes: Software for imaging (2 parts)”
organized by: Ozan Öktem (KTH - Royal Institute of Technology) , Holger Kohr (Thermo Fisher Scientific) , Jonas Adler (KTH Royal Institute of Technology) .

Authors:

Héctor Andrade Loarca (TU Berlin)

Keywords: