Brain tissue segmentation based on diffusion MRI using ℓ0 sparse-group representation classification

Pew Thian Yap, Yong Zhang, Dinggang Shen

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

3 Citations (Scopus)

Abstract

We present a method for automated brain tissue segmentation based on diffusion MRI. This provides information that is complementary to structural MRI and facilitates fusion of information between the two imaging modalities. Unlike existing segmentation approaches that are based on diffusion tensor imaging (DTI), our method explicitly models the coexistence of various diffusion compartments within each voxel owing to different tissue types and different fiber orientations. This results in improved segmentation in regions with white matter crossings and in regions susceptible to partial volume effects. For each voxel, we tease apart possible signal contributions from white matter (WM), gray matter (GM), and cerebrospinal fluid (CSF) with the help of diffusion exemplars, which are representative signals associated with each tissue type. Each voxel is then classified by determining which of the WM, GM, or CSF diffusion exemplar groups explains the signal better with the least fitting residual. Fitting is performed using ℓ0 sparse-group approximation, circumventing various reported limitations of ℓ1 fitting. In addition, to promote spatial regularity, we introduce a smoothing technique that is based on ℓ0 gradient minimization, which can be viewed as the ℓ0 version of total variation (TV) smoothing. Compared with the latter, our smoothing technique, which also incorporates multi-channelWM,GM, and CSF concurrent smoothing, yields marked improvement in preserving boundary contrast and consequently reduces segmentation bias caused by smoothing at tissue boundaries. The results produced by our method are in good agreement with segmentation based on T1-weighted images.

Original languageEnglish
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
PublisherSpringer Verlag
Pages132-139
Number of pages8
Volume9351
ISBN (Print)9783319245737
DOIs
Publication statusPublished - 2015
Externally publishedYes
Event18th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2015 - Munich, Germany
Duration: 2015 Oct 52015 Oct 9

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume9351
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other18th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2015
CountryGermany
CityMunich
Period15/10/515/10/9

Fingerprint

Group Representation
Cerebrospinal fluid
Magnetic resonance imaging
Brain
Segmentation
Tissue
Voxel
Smoothing
Smoothing Techniques
Fluid
Diffusion tensor imaging
Imaging
Fiber reinforced materials
Fiber Orientation
Total Variation
Fusion reactions
Coexistence
Modality
Concurrent
Fusion

ASJC Scopus subject areas

  • Computer Science(all)
  • Theoretical Computer Science

Cite this

Yap, P. T., Zhang, Y., & Shen, D. (2015). Brain tissue segmentation based on diffusion MRI using ℓ0 sparse-group representation classification. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 9351, pp. 132-139). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9351). Springer Verlag. https://doi.org/10.1007/978-3-319-24574-4_16

Brain tissue segmentation based on diffusion MRI using ℓ0 sparse-group representation classification. / Yap, Pew Thian; Zhang, Yong; Shen, Dinggang.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 9351 Springer Verlag, 2015. p. 132-139 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9351).

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

Yap, PT, Zhang, Y & Shen, D 2015, Brain tissue segmentation based on diffusion MRI using ℓ0 sparse-group representation classification. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). vol. 9351, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 9351, Springer Verlag, pp. 132-139, 18th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2015, Munich, Germany, 15/10/5. https://doi.org/10.1007/978-3-319-24574-4_16
Yap PT, Zhang Y, Shen D. Brain tissue segmentation based on diffusion MRI using ℓ0 sparse-group representation classification. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 9351. Springer Verlag. 2015. p. 132-139. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-319-24574-4_16
Yap, Pew Thian ; Zhang, Yong ; Shen, Dinggang. / Brain tissue segmentation based on diffusion MRI using ℓ0 sparse-group representation classification. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 9351 Springer Verlag, 2015. pp. 132-139 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{d61d6b11c9e641bf9eafde396eda82af,
title = "Brain tissue segmentation based on diffusion MRI using ℓ0 sparse-group representation classification",
abstract = "We present a method for automated brain tissue segmentation based on diffusion MRI. This provides information that is complementary to structural MRI and facilitates fusion of information between the two imaging modalities. Unlike existing segmentation approaches that are based on diffusion tensor imaging (DTI), our method explicitly models the coexistence of various diffusion compartments within each voxel owing to different tissue types and different fiber orientations. This results in improved segmentation in regions with white matter crossings and in regions susceptible to partial volume effects. For each voxel, we tease apart possible signal contributions from white matter (WM), gray matter (GM), and cerebrospinal fluid (CSF) with the help of diffusion exemplars, which are representative signals associated with each tissue type. Each voxel is then classified by determining which of the WM, GM, or CSF diffusion exemplar groups explains the signal better with the least fitting residual. Fitting is performed using ℓ0 sparse-group approximation, circumventing various reported limitations of ℓ1 fitting. In addition, to promote spatial regularity, we introduce a smoothing technique that is based on ℓ0 gradient minimization, which can be viewed as the ℓ0 version of total variation (TV) smoothing. Compared with the latter, our smoothing technique, which also incorporates multi-channelWM,GM, and CSF concurrent smoothing, yields marked improvement in preserving boundary contrast and consequently reduces segmentation bias caused by smoothing at tissue boundaries. The results produced by our method are in good agreement with segmentation based on T1-weighted images.",
author = "Yap, {Pew Thian} and Yong Zhang and Dinggang Shen",
year = "2015",
doi = "10.1007/978-3-319-24574-4_16",
language = "English",
isbn = "9783319245737",
volume = "9351",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "132--139",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

}

TY - GEN

T1 - Brain tissue segmentation based on diffusion MRI using ℓ0 sparse-group representation classification

AU - Yap, Pew Thian

AU - Zhang, Yong

AU - Shen, Dinggang

PY - 2015

Y1 - 2015

N2 - We present a method for automated brain tissue segmentation based on diffusion MRI. This provides information that is complementary to structural MRI and facilitates fusion of information between the two imaging modalities. Unlike existing segmentation approaches that are based on diffusion tensor imaging (DTI), our method explicitly models the coexistence of various diffusion compartments within each voxel owing to different tissue types and different fiber orientations. This results in improved segmentation in regions with white matter crossings and in regions susceptible to partial volume effects. For each voxel, we tease apart possible signal contributions from white matter (WM), gray matter (GM), and cerebrospinal fluid (CSF) with the help of diffusion exemplars, which are representative signals associated with each tissue type. Each voxel is then classified by determining which of the WM, GM, or CSF diffusion exemplar groups explains the signal better with the least fitting residual. Fitting is performed using ℓ0 sparse-group approximation, circumventing various reported limitations of ℓ1 fitting. In addition, to promote spatial regularity, we introduce a smoothing technique that is based on ℓ0 gradient minimization, which can be viewed as the ℓ0 version of total variation (TV) smoothing. Compared with the latter, our smoothing technique, which also incorporates multi-channelWM,GM, and CSF concurrent smoothing, yields marked improvement in preserving boundary contrast and consequently reduces segmentation bias caused by smoothing at tissue boundaries. The results produced by our method are in good agreement with segmentation based on T1-weighted images.

AB - We present a method for automated brain tissue segmentation based on diffusion MRI. This provides information that is complementary to structural MRI and facilitates fusion of information between the two imaging modalities. Unlike existing segmentation approaches that are based on diffusion tensor imaging (DTI), our method explicitly models the coexistence of various diffusion compartments within each voxel owing to different tissue types and different fiber orientations. This results in improved segmentation in regions with white matter crossings and in regions susceptible to partial volume effects. For each voxel, we tease apart possible signal contributions from white matter (WM), gray matter (GM), and cerebrospinal fluid (CSF) with the help of diffusion exemplars, which are representative signals associated with each tissue type. Each voxel is then classified by determining which of the WM, GM, or CSF diffusion exemplar groups explains the signal better with the least fitting residual. Fitting is performed using ℓ0 sparse-group approximation, circumventing various reported limitations of ℓ1 fitting. In addition, to promote spatial regularity, we introduce a smoothing technique that is based on ℓ0 gradient minimization, which can be viewed as the ℓ0 version of total variation (TV) smoothing. Compared with the latter, our smoothing technique, which also incorporates multi-channelWM,GM, and CSF concurrent smoothing, yields marked improvement in preserving boundary contrast and consequently reduces segmentation bias caused by smoothing at tissue boundaries. The results produced by our method are in good agreement with segmentation based on T1-weighted images.

UR - http://www.scopus.com/inward/record.url?scp=84951806153&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84951806153&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-24574-4_16

DO - 10.1007/978-3-319-24574-4_16

M3 - Conference contribution

AN - SCOPUS:84951806153

SN - 9783319245737

VL - 9351

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

SP - 132

EP - 139

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

PB - Springer Verlag

ER -