Receiving over 450 citations and 48000 full views, the paper entitled “Deep Convolutional Neural Network for Inverse Problems in Imaging”, appearing in IEEE Transactions on Image Processing, Volume 26, No. 9, September 2017, by Kyong Hwan Jin, Michael T. McCann, Emmanuel Froustey, and Michaël Unser has been selected for the prestigious 2019 IEEE Signal Processing Society Best Paper Award*.
Regularized iterative algorithms have emerged as the standard approach to ill-posed inverse problems in the past few decades. These methods produce excellent results, but can be challenging to deploy in practice due to factors including the high computational cost of the forward and adjoint operators and the difficulty of hyperparameter selection. Novel deep convolutional neural network (CNN)-based algorithm for solving ill-posed inverse problems is proposed. Using direct inversion followed by a CNN can solve normal-convolutional inverse problems. The direct inversion encapsulates the physical model of the system, but leads to artifacts when the problem is ill posed; the CNN combines multiresolution decomposition and residual learning in order to learn to remove these artifacts while preserving image structure.
Kyong Hwan Jin is currently a Post-Doctoral Scholar with the Biomedical Imaging Group, École Polytechnique Fédérale de Lausanne (EPFL), Lausanne, Switzerland.
Michael T. McCann was previously a postdoctoral researcher with the Biomedical Imaging Group and Research Staff Scientist working in the CIBM Signal Processing (SP) – Mathematical Imaging Section, both headed by Michaël Unser.
Emmanuel Froustey was previously a Master student with the Biomedical Imaging Group, EPFL, Lausanne, Switzerland, in 2014. He is currently a Research and Development Engineer with Dassault Aviation.
Michaël Unser, Professor at EPFL and Head of the CIBM Signal Processing (SP) – Mathematical Imaging Section, is the author of over 150 published journal papers in this area, and is one of ISI’s Highly Cited authors in Engineering. He is recipient of the IEEE Signal Processing Society’s 1995 Best Paper Award (with A. Aldroubi and M. Eden), the IEEE Signal Processing Society’s 2000 Magazine Award, and the IEEE Signal Processing Society’s 2003 Best Paper Award (with T. Blu) – making this his third IEEE Best Paper award win over the last fifteen years.
*The Best Paper Award honors the author(s) of a paper of exceptional merit dealing with a subject related to the Society’s technical scope, and appearing in one of the Society’s solely owned transactions or the Journal of Selected Topics in Signal Processing, irrespective of the author’s age.