Tight graph framelets for sparse diffusion MRI q-space representation

Pew Thian Yap, Bin Dong, Yong Zhang, Dinggang Shen

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

6 Citations (Scopus)

Abstract

In diffusion MRI,the outcome of estimation problems can often be improved by taking into account the correlation of diffusionweighted images scanned with neighboring wavevectors in q-space. For this purpose,we propose in this paper to employ tight wavelet frames constructed on non-flat domains for multi-scale sparse representation of diffusion signals. This representation is well suited for signals sampled regularly or irregularly,such as on a grid or on multiple shells,in q-space. Using spectral graph theory,the frames are constructed based on quasiaffine systems (i.e.,generalized dilations and shifts of a finite collection of wavelet functions) defined on graphs,which can be seen as a discrete representation of manifolds. The associated wavelet analysis and synthesis transforms can be computed efficiently and accurately without the need for explicit eigen-decomposition of the graph Laplacian,allowing scalability to very large problems. We demonstrate the effectiveness of this representation,generated using what we call tight graph framelets,in two specific applications: denoising and super-resolution in q-space using l0 regularization. The associated optimization problem involves only thresholding and solving a trivial inverse problem in an iterative manner. The effectiveness of graph framelets is confirmed via evaluation using synthetic data with noncentral chi noise and real data with repeated scans.

Original languageEnglish (US)
Title of host publicationMedical Image Computing and Computer-Assisted Intervention - MICCAI 2016 - 19th International Conference, Proceedings
PublisherSpringer Verlag
Pages561-569
Number of pages9
Volume9902 LNCS
ISBN (Print)9783319467252
DOIs
StatePublished - 2016
Externally publishedYes

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume9902 LNCS
ISSN (Print)03029743
ISSN (Electronic)16113349

Fingerprint

Magnetic resonance imaging
Wavelet analysis
Graph theory
Graph in graph theory
Inverse problems
Spectral Graph Theory
Graph Laplacian
Scalability
Wavelet Frames
Tight Frame
Sparse Representation
Super-resolution
Wavelet Analysis
Thresholding
Synthetic Data
Denoising
Decomposition
Dilation
Shell
Regularization

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Yap, P. T., Dong, B., Zhang, Y., & Shen, D. (2016). Tight graph framelets for sparse diffusion MRI q-space representation. In Medical Image Computing and Computer-Assisted Intervention - MICCAI 2016 - 19th International Conference, Proceedings (Vol. 9902 LNCS, pp. 561-569). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9902 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-319-46726-9_65

Tight graph framelets for sparse diffusion MRI q-space representation. / Yap, Pew Thian; Dong, Bin; Zhang, Yong; Shen, Dinggang.

Medical Image Computing and Computer-Assisted Intervention - MICCAI 2016 - 19th International Conference, Proceedings. Vol. 9902 LNCS Springer Verlag, 2016. p. 561-569 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9902 LNCS).

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

Yap, PT, Dong, B, Zhang, Y & Shen, D 2016, Tight graph framelets for sparse diffusion MRI q-space representation. in Medical Image Computing and Computer-Assisted Intervention - MICCAI 2016 - 19th International Conference, Proceedings. vol. 9902 LNCS, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 9902 LNCS, Springer Verlag, pp. 561-569. https://doi.org/10.1007/978-3-319-46726-9_65
Yap PT, Dong B, Zhang Y, Shen D. Tight graph framelets for sparse diffusion MRI q-space representation. In Medical Image Computing and Computer-Assisted Intervention - MICCAI 2016 - 19th International Conference, Proceedings. Vol. 9902 LNCS. Springer Verlag. 2016. p. 561-569. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-319-46726-9_65
Yap, Pew Thian ; Dong, Bin ; Zhang, Yong ; Shen, Dinggang. / Tight graph framelets for sparse diffusion MRI q-space representation. Medical Image Computing and Computer-Assisted Intervention - MICCAI 2016 - 19th International Conference, Proceedings. Vol. 9902 LNCS Springer Verlag, 2016. pp. 561-569 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{6ebc61115a3441f89533f9a8fcd8beae,
title = "Tight graph framelets for sparse diffusion MRI q-space representation",
abstract = "In diffusion MRI,the outcome of estimation problems can often be improved by taking into account the correlation of diffusionweighted images scanned with neighboring wavevectors in q-space. For this purpose,we propose in this paper to employ tight wavelet frames constructed on non-flat domains for multi-scale sparse representation of diffusion signals. This representation is well suited for signals sampled regularly or irregularly,such as on a grid or on multiple shells,in q-space. Using spectral graph theory,the frames are constructed based on quasiaffine systems (i.e.,generalized dilations and shifts of a finite collection of wavelet functions) defined on graphs,which can be seen as a discrete representation of manifolds. The associated wavelet analysis and synthesis transforms can be computed efficiently and accurately without the need for explicit eigen-decomposition of the graph Laplacian,allowing scalability to very large problems. We demonstrate the effectiveness of this representation,generated using what we call tight graph framelets,in two specific applications: denoising and super-resolution in q-space using l0 regularization. The associated optimization problem involves only thresholding and solving a trivial inverse problem in an iterative manner. The effectiveness of graph framelets is confirmed via evaluation using synthetic data with noncentral chi noise and real data with repeated scans.",
author = "Yap, {Pew Thian} and Bin Dong and Yong Zhang and Dinggang Shen",
year = "2016",
doi = "10.1007/978-3-319-46726-9_65",
language = "English (US)",
isbn = "9783319467252",
volume = "9902 LNCS",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "561--569",
booktitle = "Medical Image Computing and Computer-Assisted Intervention - MICCAI 2016 - 19th International Conference, Proceedings",
address = "Germany",

}

TY - GEN

T1 - Tight graph framelets for sparse diffusion MRI q-space representation

AU - Yap, Pew Thian

AU - Dong, Bin

AU - Zhang, Yong

AU - Shen, Dinggang

PY - 2016

Y1 - 2016

N2 - In diffusion MRI,the outcome of estimation problems can often be improved by taking into account the correlation of diffusionweighted images scanned with neighboring wavevectors in q-space. For this purpose,we propose in this paper to employ tight wavelet frames constructed on non-flat domains for multi-scale sparse representation of diffusion signals. This representation is well suited for signals sampled regularly or irregularly,such as on a grid or on multiple shells,in q-space. Using spectral graph theory,the frames are constructed based on quasiaffine systems (i.e.,generalized dilations and shifts of a finite collection of wavelet functions) defined on graphs,which can be seen as a discrete representation of manifolds. The associated wavelet analysis and synthesis transforms can be computed efficiently and accurately without the need for explicit eigen-decomposition of the graph Laplacian,allowing scalability to very large problems. We demonstrate the effectiveness of this representation,generated using what we call tight graph framelets,in two specific applications: denoising and super-resolution in q-space using l0 regularization. The associated optimization problem involves only thresholding and solving a trivial inverse problem in an iterative manner. The effectiveness of graph framelets is confirmed via evaluation using synthetic data with noncentral chi noise and real data with repeated scans.

AB - In diffusion MRI,the outcome of estimation problems can often be improved by taking into account the correlation of diffusionweighted images scanned with neighboring wavevectors in q-space. For this purpose,we propose in this paper to employ tight wavelet frames constructed on non-flat domains for multi-scale sparse representation of diffusion signals. This representation is well suited for signals sampled regularly or irregularly,such as on a grid or on multiple shells,in q-space. Using spectral graph theory,the frames are constructed based on quasiaffine systems (i.e.,generalized dilations and shifts of a finite collection of wavelet functions) defined on graphs,which can be seen as a discrete representation of manifolds. The associated wavelet analysis and synthesis transforms can be computed efficiently and accurately without the need for explicit eigen-decomposition of the graph Laplacian,allowing scalability to very large problems. We demonstrate the effectiveness of this representation,generated using what we call tight graph framelets,in two specific applications: denoising and super-resolution in q-space using l0 regularization. The associated optimization problem involves only thresholding and solving a trivial inverse problem in an iterative manner. The effectiveness of graph framelets is confirmed via evaluation using synthetic data with noncentral chi noise and real data with repeated scans.

UR - http://www.scopus.com/inward/record.url?scp=84996565512&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84996565512&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-46726-9_65

DO - 10.1007/978-3-319-46726-9_65

M3 - Conference contribution

AN - SCOPUS:84996565512

SN - 9783319467252

VL - 9902 LNCS

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

SP - 561

EP - 569

BT - Medical Image Computing and Computer-Assisted Intervention - MICCAI 2016 - 19th International Conference, Proceedings

PB - Springer Verlag

ER -