Machine learning techniques for image reconstructionMS65

The development of fast and accurate reconstruction algorithms is a central mathematical aspect of imaging. Most traditional image reconstruction methods can basically be classified in either analytical or iterative methods. Recently, a new class of image reconstruction methods appeared which use methods from machine learning, especially from deep learning. Initial results using deep learning techniques for image reconstruction demonstrate great promise, for example, for improving image quality, reducing computation time, or reducing radiation exposure. In this minisymposium leading experts will report on recent progress towards using machine learning for image reconstruction.

Finding best approximation pairs with Douglas-Rachford
Irene Waldspurger (CEREMADE (Université Paris-Dauphine))
Deep Learning Approaches for MR Image Reconstruction
Daniel Rueckert (Imperial College London)
Deep convolutional framelets: a general deep learning framework for inverse problems
Jong Chul Ye (Department of Bio and Brain Engineering, Korea Advanced Institute of Science and Technology)
Deep Learning for Photoacoustic Tomography from Sparse Data
Johannes Schwab (Universität Innsbruck)
Deep learning in computational microscopy
Yair Rivenson (University of California Los Angeles)
Task Based Reconstruction using Deep Learning
Ozan Öktem (KTH - Royal Institute of Technology)
Machine learning in compressed sensing
Stephan Antholzer (University of Innsbruck)
The cone-beam transform and spherical convolution operators
Michael Quellmalz (Technische Universität Chemnitz)
Markus Haltmeier (University Innsbruck)
Linh Nguyen (University of Idaho)
computed tomography, deep learning, image deblurring, image reconstruction, inverse problems, machine learning