A deep learning based anti-aliasing self super-resolution algorithm for MRI

Can Zhao, Aaron Carass, Blake E. Dewey, Jonghye Woo, Jiwon Oh, Peter Calabresi, Daniel S. Reich, Pascal Sati, Dzung L. Pham, Jerry Ladd Prince

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

High resolution magnetic resonance (MR) images are desired in many clinical applications, yet acquiring such data with an adequate signal-to-noise ratio requires a long time, making them costly and susceptible to motion artifacts. A common way to partly achieve this goal is to acquire MR images with good in-plane resolution and poor through-plane resolution (i.e., large slice thickness). For such 2D imaging protocols, aliasing is also introduced in the through-plane direction, and these high-frequency artifacts cannot be removed by conventional interpolation. Super-resolution (SR) algorithms which can reduce aliasing artifacts and improve spatial resolution have previously been reported. State-of-the-art SR methods are mostly learning-based and require external training data consisting of paired low resolution (LR) and high resolution (HR) MR images. However, due to scanner limitations, such training data are often unavailable. This paper presents an anti-aliasing (AA) and self super-resolution (SSR) algorithm that needs no external training data. It takes advantage of the fact that the in-plane slices of those MR images contain high frequency information. Our algorithm consists of three steps: (1) We build a self AA (SAA) deep network followed by (2) an SSR deep network, both of which can be applied along different orientations within the original images, and (3) recombine the multiple orientations output from Steps 1 and 2 using Fourier burst accumulation. We perform our SAA+SSR algorithm on a diverse collection of MR data without modification or preprocessing other than N4 inhomogeneity correction, and demonstrate significant improvement compared to competing SSR methods.

Original languageEnglish (US)
Title of host publicationMedical Image Computing and Computer Assisted Intervention – MICCAI 2018 - 21st International Conference, 2018, Proceedings
EditorsJulia A. Schnabel, Christos Davatzikos, Carlos Alberola-López, Gabor Fichtinger, Alejandro F. Frangi
PublisherSpringer Verlag
Pages100-108
Number of pages9
ISBN (Print)9783030009274
DOIs
StatePublished - Jan 1 2018
Event21st International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2018 - Granada, Spain
Duration: Sep 16 2018Sep 20 2018

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume11070 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other21st International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2018
CountrySpain
CityGranada
Period9/16/189/20/18

Fingerprint

Anti-aliasing
Aliasing
Super-resolution
Magnetic resonance
Magnetic resonance imaging
Magnetic Resonance Image
Slice
High Resolution
Magnetic Resonance
Scanner
Signal to noise ratio
Interpolation
Burst
Inhomogeneity
Spatial Resolution
Preprocessing
Learning
Deep learning
Imaging techniques
Interpolate

Keywords

  • Aliasing
  • CNN
  • Deep network
  • MRI
  • Self super-resolution

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Zhao, C., Carass, A., Dewey, B. E., Woo, J., Oh, J., Calabresi, P., ... Prince, J. L. (2018). A deep learning based anti-aliasing self super-resolution algorithm for MRI. In J. A. Schnabel, C. Davatzikos, C. Alberola-López, G. Fichtinger, & A. F. Frangi (Eds.), Medical Image Computing and Computer Assisted Intervention – MICCAI 2018 - 21st International Conference, 2018, Proceedings (pp. 100-108). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11070 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-030-00928-1_12

A deep learning based anti-aliasing self super-resolution algorithm for MRI. / Zhao, Can; Carass, Aaron; Dewey, Blake E.; Woo, Jonghye; Oh, Jiwon; Calabresi, Peter; Reich, Daniel S.; Sati, Pascal; Pham, Dzung L.; Prince, Jerry Ladd.

Medical Image Computing and Computer Assisted Intervention – MICCAI 2018 - 21st International Conference, 2018, Proceedings. ed. / Julia A. Schnabel; Christos Davatzikos; Carlos Alberola-López; Gabor Fichtinger; Alejandro F. Frangi. Springer Verlag, 2018. p. 100-108 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11070 LNCS).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Zhao, C, Carass, A, Dewey, BE, Woo, J, Oh, J, Calabresi, P, Reich, DS, Sati, P, Pham, DL & Prince, JL 2018, A deep learning based anti-aliasing self super-resolution algorithm for MRI. in JA Schnabel, C Davatzikos, C Alberola-López, G Fichtinger & AF Frangi (eds), Medical Image Computing and Computer Assisted Intervention – MICCAI 2018 - 21st International Conference, 2018, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 11070 LNCS, Springer Verlag, pp. 100-108, 21st International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2018, Granada, Spain, 9/16/18. https://doi.org/10.1007/978-3-030-00928-1_12
Zhao C, Carass A, Dewey BE, Woo J, Oh J, Calabresi P et al. A deep learning based anti-aliasing self super-resolution algorithm for MRI. In Schnabel JA, Davatzikos C, Alberola-López C, Fichtinger G, Frangi AF, editors, Medical Image Computing and Computer Assisted Intervention – MICCAI 2018 - 21st International Conference, 2018, Proceedings. Springer Verlag. 2018. p. 100-108. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-030-00928-1_12
Zhao, Can ; Carass, Aaron ; Dewey, Blake E. ; Woo, Jonghye ; Oh, Jiwon ; Calabresi, Peter ; Reich, Daniel S. ; Sati, Pascal ; Pham, Dzung L. ; Prince, Jerry Ladd. / A deep learning based anti-aliasing self super-resolution algorithm for MRI. Medical Image Computing and Computer Assisted Intervention – MICCAI 2018 - 21st International Conference, 2018, Proceedings. editor / Julia A. Schnabel ; Christos Davatzikos ; Carlos Alberola-López ; Gabor Fichtinger ; Alejandro F. Frangi. Springer Verlag, 2018. pp. 100-108 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{2c71d5f7f96f4f1c85624a866647ed5b,
title = "A deep learning based anti-aliasing self super-resolution algorithm for MRI",
abstract = "High resolution magnetic resonance (MR) images are desired in many clinical applications, yet acquiring such data with an adequate signal-to-noise ratio requires a long time, making them costly and susceptible to motion artifacts. A common way to partly achieve this goal is to acquire MR images with good in-plane resolution and poor through-plane resolution (i.e., large slice thickness). For such 2D imaging protocols, aliasing is also introduced in the through-plane direction, and these high-frequency artifacts cannot be removed by conventional interpolation. Super-resolution (SR) algorithms which can reduce aliasing artifacts and improve spatial resolution have previously been reported. State-of-the-art SR methods are mostly learning-based and require external training data consisting of paired low resolution (LR) and high resolution (HR) MR images. However, due to scanner limitations, such training data are often unavailable. This paper presents an anti-aliasing (AA) and self super-resolution (SSR) algorithm that needs no external training data. It takes advantage of the fact that the in-plane slices of those MR images contain high frequency information. Our algorithm consists of three steps: (1) We build a self AA (SAA) deep network followed by (2) an SSR deep network, both of which can be applied along different orientations within the original images, and (3) recombine the multiple orientations output from Steps 1 and 2 using Fourier burst accumulation. We perform our SAA+SSR algorithm on a diverse collection of MR data without modification or preprocessing other than N4 inhomogeneity correction, and demonstrate significant improvement compared to competing SSR methods.",
keywords = "Aliasing, CNN, Deep network, MRI, Self super-resolution",
author = "Can Zhao and Aaron Carass and Dewey, {Blake E.} and Jonghye Woo and Jiwon Oh and Peter Calabresi and Reich, {Daniel S.} and Pascal Sati and Pham, {Dzung L.} and Prince, {Jerry Ladd}",
year = "2018",
month = "1",
day = "1",
doi = "10.1007/978-3-030-00928-1_12",
language = "English (US)",
isbn = "9783030009274",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "100--108",
editor = "Schnabel, {Julia A.} and Christos Davatzikos and Carlos Alberola-L{\'o}pez and Gabor Fichtinger and Frangi, {Alejandro F.}",
booktitle = "Medical Image Computing and Computer Assisted Intervention – MICCAI 2018 - 21st International Conference, 2018, Proceedings",

}

TY - GEN

T1 - A deep learning based anti-aliasing self super-resolution algorithm for MRI

AU - Zhao, Can

AU - Carass, Aaron

AU - Dewey, Blake E.

AU - Woo, Jonghye

AU - Oh, Jiwon

AU - Calabresi, Peter

AU - Reich, Daniel S.

AU - Sati, Pascal

AU - Pham, Dzung L.

AU - Prince, Jerry Ladd

PY - 2018/1/1

Y1 - 2018/1/1

N2 - High resolution magnetic resonance (MR) images are desired in many clinical applications, yet acquiring such data with an adequate signal-to-noise ratio requires a long time, making them costly and susceptible to motion artifacts. A common way to partly achieve this goal is to acquire MR images with good in-plane resolution and poor through-plane resolution (i.e., large slice thickness). For such 2D imaging protocols, aliasing is also introduced in the through-plane direction, and these high-frequency artifacts cannot be removed by conventional interpolation. Super-resolution (SR) algorithms which can reduce aliasing artifacts and improve spatial resolution have previously been reported. State-of-the-art SR methods are mostly learning-based and require external training data consisting of paired low resolution (LR) and high resolution (HR) MR images. However, due to scanner limitations, such training data are often unavailable. This paper presents an anti-aliasing (AA) and self super-resolution (SSR) algorithm that needs no external training data. It takes advantage of the fact that the in-plane slices of those MR images contain high frequency information. Our algorithm consists of three steps: (1) We build a self AA (SAA) deep network followed by (2) an SSR deep network, both of which can be applied along different orientations within the original images, and (3) recombine the multiple orientations output from Steps 1 and 2 using Fourier burst accumulation. We perform our SAA+SSR algorithm on a diverse collection of MR data without modification or preprocessing other than N4 inhomogeneity correction, and demonstrate significant improvement compared to competing SSR methods.

AB - High resolution magnetic resonance (MR) images are desired in many clinical applications, yet acquiring such data with an adequate signal-to-noise ratio requires a long time, making them costly and susceptible to motion artifacts. A common way to partly achieve this goal is to acquire MR images with good in-plane resolution and poor through-plane resolution (i.e., large slice thickness). For such 2D imaging protocols, aliasing is also introduced in the through-plane direction, and these high-frequency artifacts cannot be removed by conventional interpolation. Super-resolution (SR) algorithms which can reduce aliasing artifacts and improve spatial resolution have previously been reported. State-of-the-art SR methods are mostly learning-based and require external training data consisting of paired low resolution (LR) and high resolution (HR) MR images. However, due to scanner limitations, such training data are often unavailable. This paper presents an anti-aliasing (AA) and self super-resolution (SSR) algorithm that needs no external training data. It takes advantage of the fact that the in-plane slices of those MR images contain high frequency information. Our algorithm consists of three steps: (1) We build a self AA (SAA) deep network followed by (2) an SSR deep network, both of which can be applied along different orientations within the original images, and (3) recombine the multiple orientations output from Steps 1 and 2 using Fourier burst accumulation. We perform our SAA+SSR algorithm on a diverse collection of MR data without modification or preprocessing other than N4 inhomogeneity correction, and demonstrate significant improvement compared to competing SSR methods.

KW - Aliasing

KW - CNN

KW - Deep network

KW - MRI

KW - Self super-resolution

UR - http://www.scopus.com/inward/record.url?scp=85054053751&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85054053751&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-00928-1_12

DO - 10.1007/978-3-030-00928-1_12

M3 - Conference contribution

AN - SCOPUS:85054053751

SN - 9783030009274

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

SP - 100

EP - 108

BT - Medical Image Computing and Computer Assisted Intervention – MICCAI 2018 - 21st International Conference, 2018, Proceedings

A2 - Schnabel, Julia A.

A2 - Davatzikos, Christos

A2 - Alberola-López, Carlos

A2 - Fichtinger, Gabor

A2 - Frangi, Alejandro F.

PB - Springer Verlag

ER -