Ändra sökning
RefereraExporteraLänk till posten
Permanent länk

Direktlänk
Referera
Referensformat
  • apa
  • harvard1
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Annat format
Fler format
Språk
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Annat språk
Fler språk
Utmatningsformat
  • html
  • text
  • asciidoc
  • rtf
Solving ill-posed inverse problems using iterative deep neural networks
KTH, Skolan för teknikvetenskap (SCI), Matematik (Inst.), Matematik (Avd.).
KTH, Skolan för teknikvetenskap (SCI), Matematik (Inst.), Matematik (Avd.).ORCID-id: 0000-0002-1118-6483
2017 (Engelska)Ingår i: Inverse Problems, ISSN 0266-5611, E-ISSN 1361-6420, Vol. 33, nr 12, artikel-id 124007Artikel i tidskrift (Refereegranskat) Published
Abstract [en]

We propose a partially learned approach for the solution of ill-posed inverse problems with not necessarily linear forward operators. The method builds on ideas from classical regularisation theory and recent advances in deep learning to perform learning while making use of prior information about the inverse problem encoded in the forward operator, noise model and a regularising functional. The method results in a gradient-like iterative scheme, where the 'gradient' component is learned using a convolutional network that includes the gradients of the data discrepancy and regulariser as input in each iteration. We present results of such a partially learned gradient scheme on a non-linear tomographic inversion problem with simulated data from both the Sheep-Logan phantom as well as a head CT. The outcome is compared against filtered backprojection and total variation reconstruction and the proposed method provides a 5.4 dB PSNR improvement over the total variation reconstruction while being significantly faster, giving reconstructions of 512 x 512 pixel images in about 0.4 s using a single graphics processing unit (GPU).

Ort, förlag, år, upplaga, sidor
Institute of Physics Publishing (IOPP), 2017. Vol. 33, nr 12, artikel-id 124007
Nyckelord [en]
tomography, deep learning, gradient descent, regularization
Nationell ämneskategori
Matematik
Identifikatorer
URN: urn:nbn:se:kth:diva-219496DOI: 10.1088/1361-6420/aa9581ISI: 000416015300001Scopus ID: 2-s2.0-85038424472OAI: oai:DiVA.org:kth-219496DiVA, id: diva2:1163554
Anmärkning

QC 20171207

Tillgänglig från: 2017-12-07 Skapad: 2017-12-07 Senast uppdaterad: 2019-10-18Bibliografiskt granskad
Ingår i avhandling
1. Data-driven Methods in Inverse Problems
Öppna denna publikation i ny flik eller fönster >>Data-driven Methods in Inverse Problems
2019 (Engelska)Doktorsavhandling, sammanläggning (Övrigt vetenskapligt)
Abstract [en]

In this thesis on data-driven methods in inverse problems we introduce several new methods to solve inverse problems using recent advancements in machine learning and specifically deep learning. The main goal has been to develop practically applicable methods, scalable to medical applications and with the ability to handle all the complexities associated with them.

In total, the thesis contains six papers. Some of them are focused on more theoretical questions such as characterizing the optimal solutions of reconstruction schemes or extending current methods to new domains, while others have focused on practical applicability. A significant portion of the papers also aim to bringing knowledge from the machine learning community into the imaging community, with considerable effort spent on translating many of the concepts. The papers have been published in a range of venues: machine learning, medical imaging and inverse problems.

The first two papers contribute to a class of methods now called learned iterative reconstruction where we introduce two ways of combining classical model driven reconstruction methods with deep neural networks. The next two papers look forward, aiming to address the question of "what do we want?" by proposing two very different but novel loss functions for training neural networks in inverse problems. The final papers dwelve into the statistical side, one gives a generalization of a class of deep generative models to Banach spaces while the next introduces two ways in which such methods can be used to perform Bayesian inversion at scale.

Ort, förlag, år, upplaga, sidor
Stockholm: KTH Royal Institute of Technology, 2019. s. 196
Serie
TRITA-SCI-FOU ; 2019;49
Nyckelord
Inverse Problems, Machine Learning, Tomography
Nationell ämneskategori
Beräkningsmatematik
Identifikatorer
urn:nbn:se:kth:diva-262727 (URN)978-91-7873-334-7 (ISBN)
Disputation
2019-10-31, F3, Lindstedtsvägen26, KTH, Stockholm, 14:00 (Engelska)
Opponent
Handledare
Forskningsfinansiär
Stiftelsen för strategisk forskning (SSF)
Tillgänglig från: 2019-10-21 Skapad: 2019-10-18 Senast uppdaterad: 2019-10-21Bibliografiskt granskad

Open Access i DiVA

Fulltext saknas i DiVA

Övriga länkar

Förlagets fulltextScopus

Personposter BETA

Öktem, Ozan

Sök vidare i DiVA

Av författaren/redaktören
Adler, JonasÖktem, Ozan
Av organisationen
Matematik (Avd.)
I samma tidskrift
Inverse Problems
Matematik

Sök vidare utanför DiVA

GoogleGoogle Scholar

doi
urn-nbn

Altmetricpoäng

doi
urn-nbn
Totalt: 2215 träffar
RefereraExporteraLänk till posten
Permanent länk

Direktlänk
Referera
Referensformat
  • apa
  • harvard1
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Annat format
Fler format
Språk
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Annat språk
Fler språk
Utmatningsformat
  • html
  • text
  • asciidoc
  • rtf