Parareal Method for Anisotropic Diffusion Denoising

Xiujie Shan; Martin B. van Gijzen

doi:10.1007/978-3-031-30445-3_26

Parareal Method for Anisotropic Diffusion Denoising

^*Corresponding author for this work

Numerical Analysis

Research output: Chapter in Book/Conference proceedings/Edited volume › Conference contribution › Scientific › peer-review

21 Downloads (Pure)

Abstract

This paper studies time-domain parallelisation using Parareal to speed up the computations of anisotropic diffusion filtering. We consider both explicit and implicit Euler based method for the propagation in time for Parareal. The Preconditioned Conjugate Gradient (PCG) method is used to solve the systems that arise in the implicit based method. The estimation of the iteration numbers of PCG allows us to predict the running time of Parareal calculation, which further guides us in the experimental stage. Parallelisation of the method is implemented using Coarray Fortran. We illustrate the experimental results on 3D low-field MRI images using up to 960 cores. The computational improvement in time is achieved.

Original language	English
Title of host publication	Parallel Processing and Applied Mathematics - 14th International Conference, PPAM 2022, Revised Selected Papers
Editors	Roman Wyrzykowski, Jack Dongarra, Ewa Deelman, Konrad Karczewski
Publisher	Springer
Pages	313-322
Number of pages	10
ISBN (Print)	9783031304446
DOIs	https://doi.org/10.1007/978-3-031-30445-3_26
Publication status	Published - 2023
Event	14th International Conference on Parallel Processing and Applied Mathematics, PPAM 2022 - Gdansk, Poland Duration: 11 Sept 2022 → 14 Sept 2022

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	13827 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	14th International Conference on Parallel Processing and Applied Mathematics, PPAM 2022
Country/Territory	Poland
City	Gdansk
Period	11/09/22 → 14/09/22

Bibliographical note

Green Open Access added to TU Delft Institutional Repository ‘You share, we take care!’ – Taverne project https://www.openaccess.nl/en/you-share-we-take-care
Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public.

Keywords

High performance computing
Image denoising
Nonlinear Diffusion equation
Parareal

Access to Document

10.1007/978-3-031-30445-3_26

978-3-031-30445-3_26Final published version, 779 KB

Cite this

Shan, X., & van Gijzen, M. B. (2023). Parareal Method for Anisotropic Diffusion Denoising. In R. Wyrzykowski, J. Dongarra, E. Deelman, & K. Karczewski (Eds.), Parallel Processing and Applied Mathematics - 14th International Conference, PPAM 2022, Revised Selected Papers (pp. 313-322). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 13827 LNCS). Springer. https://doi.org/10.1007/978-3-031-30445-3_26

Shan, Xiujie ; van Gijzen, Martin B. / Parareal Method for Anisotropic Diffusion Denoising. Parallel Processing and Applied Mathematics - 14th International Conference, PPAM 2022, Revised Selected Papers. editor / Roman Wyrzykowski ; Jack Dongarra ; Ewa Deelman ; Konrad Karczewski. Springer, 2023. pp. 313-322 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{f19fba4275824b15acee7c015db68c91,

title = "Parareal Method for Anisotropic Diffusion Denoising",

abstract = "This paper studies time-domain parallelisation using Parareal to speed up the computations of anisotropic diffusion filtering. We consider both explicit and implicit Euler based method for the propagation in time for Parareal. The Preconditioned Conjugate Gradient (PCG) method is used to solve the systems that arise in the implicit based method. The estimation of the iteration numbers of PCG allows us to predict the running time of Parareal calculation, which further guides us in the experimental stage. Parallelisation of the method is implemented using Coarray Fortran. We illustrate the experimental results on 3D low-field MRI images using up to 960 cores. The computational improvement in time is achieved.",

keywords = "High performance computing, Image denoising, Nonlinear Diffusion equation, Parareal",

author = "Xiujie Shan and {van Gijzen}, {Martin B.}",

note = "Green Open Access added to TU Delft Institutional Repository {\textquoteleft}You share, we take care!{\textquoteright} – Taverne project https://www.openaccess.nl/en/you-share-we-take-care Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public. ; 14th International Conference on Parallel Processing and Applied Mathematics, PPAM 2022 ; Conference date: 11-09-2022 Through 14-09-2022",

year = "2023",

doi = "10.1007/978-3-031-30445-3_26",

language = "English",

isbn = "9783031304446",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer",

pages = "313--322",

editor = "Roman Wyrzykowski and Jack Dongarra and Ewa Deelman and Konrad Karczewski",

booktitle = "Parallel Processing and Applied Mathematics - 14th International Conference, PPAM 2022, Revised Selected Papers",

}

Shan, X & van Gijzen, MB 2023, Parareal Method for Anisotropic Diffusion Denoising. in R Wyrzykowski, J Dongarra, E Deelman & K Karczewski (eds), Parallel Processing and Applied Mathematics - 14th International Conference, PPAM 2022, Revised Selected Papers. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 13827 LNCS, Springer, pp. 313-322, 14th International Conference on Parallel Processing and Applied Mathematics, PPAM 2022, Gdansk, Poland, 11/09/22. https://doi.org/10.1007/978-3-031-30445-3_26

Parareal Method for Anisotropic Diffusion Denoising. / Shan, Xiujie; van Gijzen, Martin B.
Parallel Processing and Applied Mathematics - 14th International Conference, PPAM 2022, Revised Selected Papers. ed. / Roman Wyrzykowski; Jack Dongarra; Ewa Deelman; Konrad Karczewski. Springer, 2023. p. 313-322 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 13827 LNCS).

Research output: Chapter in Book/Conference proceedings/Edited volume › Conference contribution › Scientific › peer-review

TY - GEN

T1 - Parareal Method for Anisotropic Diffusion Denoising

AU - Shan, Xiujie

AU - van Gijzen, Martin B.

N1 - Green Open Access added to TU Delft Institutional Repository ‘You share, we take care!’ – Taverne project https://www.openaccess.nl/en/you-share-we-take-care Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public.

PY - 2023

Y1 - 2023

N2 - This paper studies time-domain parallelisation using Parareal to speed up the computations of anisotropic diffusion filtering. We consider both explicit and implicit Euler based method for the propagation in time for Parareal. The Preconditioned Conjugate Gradient (PCG) method is used to solve the systems that arise in the implicit based method. The estimation of the iteration numbers of PCG allows us to predict the running time of Parareal calculation, which further guides us in the experimental stage. Parallelisation of the method is implemented using Coarray Fortran. We illustrate the experimental results on 3D low-field MRI images using up to 960 cores. The computational improvement in time is achieved.

AB - This paper studies time-domain parallelisation using Parareal to speed up the computations of anisotropic diffusion filtering. We consider both explicit and implicit Euler based method for the propagation in time for Parareal. The Preconditioned Conjugate Gradient (PCG) method is used to solve the systems that arise in the implicit based method. The estimation of the iteration numbers of PCG allows us to predict the running time of Parareal calculation, which further guides us in the experimental stage. Parallelisation of the method is implemented using Coarray Fortran. We illustrate the experimental results on 3D low-field MRI images using up to 960 cores. The computational improvement in time is achieved.

KW - High performance computing

KW - Image denoising

KW - Nonlinear Diffusion equation

KW - Parareal

UR - http://www.scopus.com/inward/record.url?scp=85161378436&partnerID=8YFLogxK

U2 - 10.1007/978-3-031-30445-3_26

DO - 10.1007/978-3-031-30445-3_26

M3 - Conference contribution

AN - SCOPUS:85161378436

SN - 9783031304446

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 313

EP - 322

BT - Parallel Processing and Applied Mathematics - 14th International Conference, PPAM 2022, Revised Selected Papers

A2 - Wyrzykowski, Roman

A2 - Dongarra, Jack

A2 - Deelman, Ewa

A2 - Karczewski, Konrad

PB - Springer

T2 - 14th International Conference on Parallel Processing and Applied Mathematics, PPAM 2022

Y2 - 11 September 2022 through 14 September 2022

ER -

Shan X, van Gijzen MB. Parareal Method for Anisotropic Diffusion Denoising. In Wyrzykowski R, Dongarra J, Deelman E, Karczewski K, editors, Parallel Processing and Applied Mathematics - 14th International Conference, PPAM 2022, Revised Selected Papers. Springer. 2023. p. 313-322. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-031-30445-3_26