Image super-resolution by supervised adaption of patchwise self-similarity from high-resolution image

Guorong Wu, Xiaofeng Zhu, Qian Wang, Dinggang Shen

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

2 Citations (Scopus)

Abstract

Image super-resolution is of great interest in medical imaging field. However, different from natural images studied in computer vision field, the low-resolution (LR) medical imaging data is often a stack of high-resolution (HR) 2D slices with large slice thickness. Consequently, the goal of superresolution for medical imaging data is to reconstruct the missing slice(s) between any two consecutive slices. Since some modalities (e.g., T1-weighted MR image) are often acquired with high-resolution (HR) image, it is intuitive to harness the prior self-similarity information in the HR image for guiding the super-resolution of LR image (e.g., T2-weighted MR image). The conventional way is to find the profile of patchwise self-similarity in the HR image and then use it to reconstruct the missing information at the same location of LR image. However, the local morphological patterns could vary significantly across the LR and HR images, due to the use of different imaging protocols. Therefore, such direct (un-supervised) adaption of self-similarity profile from HR image is often not effective in revealing the actual information in the LR image. To this end, we propose to employ the existing image information in the LR image to supervise the estimation of self-similarity profile by requiring it not only being optimal in representing patches in the HR image, but also producing less reconstruction errors for the existing image information in the LR image. Moreover, to make the anatomical structures spatially consistent in the reconstructed image, we simultaneously estimate the self-similarity profiles for a stack of patches across consecutive slices by solving a group sparse patch representation problem. We have evaluated our proposed super-resolution method on both simulated brain MR images and real patient images with multiple sclerosis lesion, achieving promising results with more anatomical details and sharpness.

Original languageEnglish
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
PublisherSpringer Verlag
Pages10-18
Number of pages9
Volume9467
ISBN (Print)9783319281933
DOIs
Publication statusPublished - 2015
Externally publishedYes
Event1st International Workshop on Patch-Based Techniques in Medical Imaging, Patch-MI 2015 - Munich, Germany
Duration: 2015 Oct 92015 Oct 9

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume9467
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other1st International Workshop on Patch-Based Techniques in Medical Imaging, Patch-MI 2015
CountryGermany
CityMunich
Period15/10/915/10/9

Fingerprint

Super-resolution
Self-similarity
Image resolution
High Resolution
Medical imaging
Slice
Optical resolving power
Medical Imaging
Patch
Consecutive
Computer vision
Brain
Multiple Sclerosis
Imaging techniques

ASJC Scopus subject areas

  • Computer Science(all)
  • Theoretical Computer Science

Cite this

Wu, G., Zhu, X., Wang, Q., & Shen, D. (2015). Image super-resolution by supervised adaption of patchwise self-similarity from high-resolution image. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 9467, pp. 10-18). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9467). Springer Verlag. https://doi.org/10.1007/978-3-319-28194-0_2

Image super-resolution by supervised adaption of patchwise self-similarity from high-resolution image. / Wu, Guorong; Zhu, Xiaofeng; Wang, Qian; Shen, Dinggang.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 9467 Springer Verlag, 2015. p. 10-18 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9467).

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

Wu, G, Zhu, X, Wang, Q & Shen, D 2015, Image super-resolution by supervised adaption of patchwise self-similarity from high-resolution image. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). vol. 9467, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 9467, Springer Verlag, pp. 10-18, 1st International Workshop on Patch-Based Techniques in Medical Imaging, Patch-MI 2015, Munich, Germany, 15/10/9. https://doi.org/10.1007/978-3-319-28194-0_2
Wu G, Zhu X, Wang Q, Shen D. Image super-resolution by supervised adaption of patchwise self-similarity from high-resolution image. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 9467. Springer Verlag. 2015. p. 10-18. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-319-28194-0_2
Wu, Guorong ; Zhu, Xiaofeng ; Wang, Qian ; Shen, Dinggang. / Image super-resolution by supervised adaption of patchwise self-similarity from high-resolution image. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 9467 Springer Verlag, 2015. pp. 10-18 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{830eaa1a8b7d4fcaa746b9eeaef7f4c8,
title = "Image super-resolution by supervised adaption of patchwise self-similarity from high-resolution image",
abstract = "Image super-resolution is of great interest in medical imaging field. However, different from natural images studied in computer vision field, the low-resolution (LR) medical imaging data is often a stack of high-resolution (HR) 2D slices with large slice thickness. Consequently, the goal of superresolution for medical imaging data is to reconstruct the missing slice(s) between any two consecutive slices. Since some modalities (e.g., T1-weighted MR image) are often acquired with high-resolution (HR) image, it is intuitive to harness the prior self-similarity information in the HR image for guiding the super-resolution of LR image (e.g., T2-weighted MR image). The conventional way is to find the profile of patchwise self-similarity in the HR image and then use it to reconstruct the missing information at the same location of LR image. However, the local morphological patterns could vary significantly across the LR and HR images, due to the use of different imaging protocols. Therefore, such direct (un-supervised) adaption of self-similarity profile from HR image is often not effective in revealing the actual information in the LR image. To this end, we propose to employ the existing image information in the LR image to supervise the estimation of self-similarity profile by requiring it not only being optimal in representing patches in the HR image, but also producing less reconstruction errors for the existing image information in the LR image. Moreover, to make the anatomical structures spatially consistent in the reconstructed image, we simultaneously estimate the self-similarity profiles for a stack of patches across consecutive slices by solving a group sparse patch representation problem. We have evaluated our proposed super-resolution method on both simulated brain MR images and real patient images with multiple sclerosis lesion, achieving promising results with more anatomical details and sharpness.",
author = "Guorong Wu and Xiaofeng Zhu and Qian Wang and Dinggang Shen",
year = "2015",
doi = "10.1007/978-3-319-28194-0_2",
language = "English",
isbn = "9783319281933",
volume = "9467",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "10--18",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

}

TY - GEN

T1 - Image super-resolution by supervised adaption of patchwise self-similarity from high-resolution image

AU - Wu, Guorong

AU - Zhu, Xiaofeng

AU - Wang, Qian

AU - Shen, Dinggang

PY - 2015

Y1 - 2015

N2 - Image super-resolution is of great interest in medical imaging field. However, different from natural images studied in computer vision field, the low-resolution (LR) medical imaging data is often a stack of high-resolution (HR) 2D slices with large slice thickness. Consequently, the goal of superresolution for medical imaging data is to reconstruct the missing slice(s) between any two consecutive slices. Since some modalities (e.g., T1-weighted MR image) are often acquired with high-resolution (HR) image, it is intuitive to harness the prior self-similarity information in the HR image for guiding the super-resolution of LR image (e.g., T2-weighted MR image). The conventional way is to find the profile of patchwise self-similarity in the HR image and then use it to reconstruct the missing information at the same location of LR image. However, the local morphological patterns could vary significantly across the LR and HR images, due to the use of different imaging protocols. Therefore, such direct (un-supervised) adaption of self-similarity profile from HR image is often not effective in revealing the actual information in the LR image. To this end, we propose to employ the existing image information in the LR image to supervise the estimation of self-similarity profile by requiring it not only being optimal in representing patches in the HR image, but also producing less reconstruction errors for the existing image information in the LR image. Moreover, to make the anatomical structures spatially consistent in the reconstructed image, we simultaneously estimate the self-similarity profiles for a stack of patches across consecutive slices by solving a group sparse patch representation problem. We have evaluated our proposed super-resolution method on both simulated brain MR images and real patient images with multiple sclerosis lesion, achieving promising results with more anatomical details and sharpness.

AB - Image super-resolution is of great interest in medical imaging field. However, different from natural images studied in computer vision field, the low-resolution (LR) medical imaging data is often a stack of high-resolution (HR) 2D slices with large slice thickness. Consequently, the goal of superresolution for medical imaging data is to reconstruct the missing slice(s) between any two consecutive slices. Since some modalities (e.g., T1-weighted MR image) are often acquired with high-resolution (HR) image, it is intuitive to harness the prior self-similarity information in the HR image for guiding the super-resolution of LR image (e.g., T2-weighted MR image). The conventional way is to find the profile of patchwise self-similarity in the HR image and then use it to reconstruct the missing information at the same location of LR image. However, the local morphological patterns could vary significantly across the LR and HR images, due to the use of different imaging protocols. Therefore, such direct (un-supervised) adaption of self-similarity profile from HR image is often not effective in revealing the actual information in the LR image. To this end, we propose to employ the existing image information in the LR image to supervise the estimation of self-similarity profile by requiring it not only being optimal in representing patches in the HR image, but also producing less reconstruction errors for the existing image information in the LR image. Moreover, to make the anatomical structures spatially consistent in the reconstructed image, we simultaneously estimate the self-similarity profiles for a stack of patches across consecutive slices by solving a group sparse patch representation problem. We have evaluated our proposed super-resolution method on both simulated brain MR images and real patient images with multiple sclerosis lesion, achieving promising results with more anatomical details and sharpness.

UR - http://www.scopus.com/inward/record.url?scp=84955309048&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84955309048&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-28194-0_2

DO - 10.1007/978-3-319-28194-0_2

M3 - Conference contribution

C2 - 30294727

AN - SCOPUS:84955309048

SN - 9783319281933

VL - 9467

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

SP - 10

EP - 18

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

PB - Springer Verlag

ER -