Recent work in machine learning shows that deep neural networks can be used
to solve a wide variety of inverse problems arising in computational imaging.
We explore the central prevailing themes of this emerging area and present a
taxonomy that can be used to categorize different problems and reconstruction
methods. Our taxonomy is organized along two central axes: (1) whether or not a
forward model is known and to what extent it is used in training and testing,
and (2) whether or not the learning is supervised or unsupervised, i.e.,
whether or not the training relies on access to matched ground truth image and
measurement pairs. We also discuss the trade-offs associated with these
different reconstruction approaches, caveats and common failure modes, plus
open problems and avenues for future work.
Описание
[2005.06001] Deep Learning Techniques for Inverse Problems in Imaging
%0 Journal Article
%1 ongie2020learning
%A Ongie, Gregory
%A Jalal, Ajil
%A Metzler, Christopher A.
%A Baraniuk, Richard G.
%A Dimakis, Alexandros G.
%A Willett, Rebecca
%D 2020
%K deep-learning survey
%T Deep Learning Techniques for Inverse Problems in Imaging
%U http://arxiv.org/abs/2005.06001
%X Recent work in machine learning shows that deep neural networks can be used
to solve a wide variety of inverse problems arising in computational imaging.
We explore the central prevailing themes of this emerging area and present a
taxonomy that can be used to categorize different problems and reconstruction
methods. Our taxonomy is organized along two central axes: (1) whether or not a
forward model is known and to what extent it is used in training and testing,
and (2) whether or not the learning is supervised or unsupervised, i.e.,
whether or not the training relies on access to matched ground truth image and
measurement pairs. We also discuss the trade-offs associated with these
different reconstruction approaches, caveats and common failure modes, plus
open problems and avenues for future work.
@article{ongie2020learning,
abstract = {Recent work in machine learning shows that deep neural networks can be used
to solve a wide variety of inverse problems arising in computational imaging.
We explore the central prevailing themes of this emerging area and present a
taxonomy that can be used to categorize different problems and reconstruction
methods. Our taxonomy is organized along two central axes: (1) whether or not a
forward model is known and to what extent it is used in training and testing,
and (2) whether or not the learning is supervised or unsupervised, i.e.,
whether or not the training relies on access to matched ground truth image and
measurement pairs. We also discuss the trade-offs associated with these
different reconstruction approaches, caveats and common failure modes, plus
open problems and avenues for future work.},
added-at = {2020-05-14T22:19:58.000+0200},
author = {Ongie, Gregory and Jalal, Ajil and Metzler, Christopher A. and Baraniuk, Richard G. and Dimakis, Alexandros G. and Willett, Rebecca},
biburl = {https://www.bibsonomy.org/bibtex/20eb7e824533fa884032d67496b3830b2/kirk86},
description = {[2005.06001] Deep Learning Techniques for Inverse Problems in Imaging},
interhash = {b1075a97bcd9b3521ca54876293f6bce},
intrahash = {0eb7e824533fa884032d67496b3830b2},
keywords = {deep-learning survey},
note = {cite arxiv:2005.06001},
timestamp = {2020-05-14T22:19:58.000+0200},
title = {Deep Learning Techniques for Inverse Problems in Imaging},
url = {http://arxiv.org/abs/2005.06001},
year = 2020
}