Segmentation of perivascular spaces using vascular features and structured random forest from 7T MR image

Jun Zhang, Yaozong Gao, Sang Hyun Park, Xiaopeng Zong, Weili Lin, Dinggang Shen

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

3 Citations (Scopus)

Abstract

Quantitative analysis of perivascular spaces (PVSs) is important to reveal the correlations between cerebrovascular lesions and neurodegenerative diseases. In this study, we propose a learning-based segmentation framework to extract the PVSs from high-resolution 7T MR images. Specifically, we integrate three types of vascular filter responses into a structured random forest for classifying voxels into PVS and background. In addition, we also propose a novel entropy-based sampling strategy to extract informative samples in the background for training the classification model. Since various vascular features can be extracted by the three vascular filters, even thin and low-contrast structures can be effectively extracted from the noisy background. Moreover, continuous and smooth segmentation results can be obtained by utilizing the patch-based structured labels. The segmentation performance is evaluated on 19 subjects with 7T MR images, and the experimental results demonstrate that the joint use of entropy-based sampling strategy, vascular features and structured learning improves the segmentation accuracy, with the Dice similarity coefficient reaching 66%.

Original languageEnglish
Title of host publicationMachine Learning in Medical Imaging - 7th International Workshop, MLMI 2016 held in conjunction with MICCAI 2016, Proceedings
PublisherSpringer Verlag
Pages61-68
Number of pages8
Volume10019 LNCS
ISBN (Print)9783319471563
DOIs
Publication statusPublished - 2016
Event7th International Workshop on Machine Learning in Medical Imaging, MLMI 2016 held in conjunction with 19th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2016 - Athens, Greece
Duration: 2016 Oct 172016 Oct 17

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume10019 LNCS
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other7th International Workshop on Machine Learning in Medical Imaging, MLMI 2016 held in conjunction with 19th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2016
CountryGreece
CityAthens
Period16/10/1716/10/17

Fingerprint

Random Forest
Entropy
Segmentation
Neurodegenerative diseases
Sampling
Sampling Strategy
Labels
Similarity Coefficient
Filter
Dice
Voxel
Quantitative Analysis
Chemical analysis
Patch
High Resolution
Integrate
Experimental Results
Demonstrate
Background
Learning

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Zhang, J., Gao, Y., Park, S. H., Zong, X., Lin, W., & Shen, D. (2016). Segmentation of perivascular spaces using vascular features and structured random forest from 7T MR image. In Machine Learning in Medical Imaging - 7th International Workshop, MLMI 2016 held in conjunction with MICCAI 2016, Proceedings (Vol. 10019 LNCS, pp. 61-68). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10019 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-319-47157-0_8

Segmentation of perivascular spaces using vascular features and structured random forest from 7T MR image. / Zhang, Jun; Gao, Yaozong; Park, Sang Hyun; Zong, Xiaopeng; Lin, Weili; Shen, Dinggang.

Machine Learning in Medical Imaging - 7th International Workshop, MLMI 2016 held in conjunction with MICCAI 2016, Proceedings. Vol. 10019 LNCS Springer Verlag, 2016. p. 61-68 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10019 LNCS).

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

Zhang, J, Gao, Y, Park, SH, Zong, X, Lin, W & Shen, D 2016, Segmentation of perivascular spaces using vascular features and structured random forest from 7T MR image. in Machine Learning in Medical Imaging - 7th International Workshop, MLMI 2016 held in conjunction with MICCAI 2016, Proceedings. vol. 10019 LNCS, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 10019 LNCS, Springer Verlag, pp. 61-68, 7th International Workshop on Machine Learning in Medical Imaging, MLMI 2016 held in conjunction with 19th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2016, Athens, Greece, 16/10/17. https://doi.org/10.1007/978-3-319-47157-0_8
Zhang J, Gao Y, Park SH, Zong X, Lin W, Shen D. Segmentation of perivascular spaces using vascular features and structured random forest from 7T MR image. In Machine Learning in Medical Imaging - 7th International Workshop, MLMI 2016 held in conjunction with MICCAI 2016, Proceedings. Vol. 10019 LNCS. Springer Verlag. 2016. p. 61-68. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-319-47157-0_8
Zhang, Jun ; Gao, Yaozong ; Park, Sang Hyun ; Zong, Xiaopeng ; Lin, Weili ; Shen, Dinggang. / Segmentation of perivascular spaces using vascular features and structured random forest from 7T MR image. Machine Learning in Medical Imaging - 7th International Workshop, MLMI 2016 held in conjunction with MICCAI 2016, Proceedings. Vol. 10019 LNCS Springer Verlag, 2016. pp. 61-68 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{0c6a0afdfa644db08bd0aad7b56a465c,
title = "Segmentation of perivascular spaces using vascular features and structured random forest from 7T MR image",
abstract = "Quantitative analysis of perivascular spaces (PVSs) is important to reveal the correlations between cerebrovascular lesions and neurodegenerative diseases. In this study, we propose a learning-based segmentation framework to extract the PVSs from high-resolution 7T MR images. Specifically, we integrate three types of vascular filter responses into a structured random forest for classifying voxels into PVS and background. In addition, we also propose a novel entropy-based sampling strategy to extract informative samples in the background for training the classification model. Since various vascular features can be extracted by the three vascular filters, even thin and low-contrast structures can be effectively extracted from the noisy background. Moreover, continuous and smooth segmentation results can be obtained by utilizing the patch-based structured labels. The segmentation performance is evaluated on 19 subjects with 7T MR images, and the experimental results demonstrate that the joint use of entropy-based sampling strategy, vascular features and structured learning improves the segmentation accuracy, with the Dice similarity coefficient reaching 66{\%}.",
author = "Jun Zhang and Yaozong Gao and Park, {Sang Hyun} and Xiaopeng Zong and Weili Lin and Dinggang Shen",
year = "2016",
doi = "10.1007/978-3-319-47157-0_8",
language = "English",
isbn = "9783319471563",
volume = "10019 LNCS",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "61--68",
booktitle = "Machine Learning in Medical Imaging - 7th International Workshop, MLMI 2016 held in conjunction with MICCAI 2016, Proceedings",

}

TY - GEN

T1 - Segmentation of perivascular spaces using vascular features and structured random forest from 7T MR image

AU - Zhang, Jun

AU - Gao, Yaozong

AU - Park, Sang Hyun

AU - Zong, Xiaopeng

AU - Lin, Weili

AU - Shen, Dinggang

PY - 2016

Y1 - 2016

N2 - Quantitative analysis of perivascular spaces (PVSs) is important to reveal the correlations between cerebrovascular lesions and neurodegenerative diseases. In this study, we propose a learning-based segmentation framework to extract the PVSs from high-resolution 7T MR images. Specifically, we integrate three types of vascular filter responses into a structured random forest for classifying voxels into PVS and background. In addition, we also propose a novel entropy-based sampling strategy to extract informative samples in the background for training the classification model. Since various vascular features can be extracted by the three vascular filters, even thin and low-contrast structures can be effectively extracted from the noisy background. Moreover, continuous and smooth segmentation results can be obtained by utilizing the patch-based structured labels. The segmentation performance is evaluated on 19 subjects with 7T MR images, and the experimental results demonstrate that the joint use of entropy-based sampling strategy, vascular features and structured learning improves the segmentation accuracy, with the Dice similarity coefficient reaching 66%.

AB - Quantitative analysis of perivascular spaces (PVSs) is important to reveal the correlations between cerebrovascular lesions and neurodegenerative diseases. In this study, we propose a learning-based segmentation framework to extract the PVSs from high-resolution 7T MR images. Specifically, we integrate three types of vascular filter responses into a structured random forest for classifying voxels into PVS and background. In addition, we also propose a novel entropy-based sampling strategy to extract informative samples in the background for training the classification model. Since various vascular features can be extracted by the three vascular filters, even thin and low-contrast structures can be effectively extracted from the noisy background. Moreover, continuous and smooth segmentation results can be obtained by utilizing the patch-based structured labels. The segmentation performance is evaluated on 19 subjects with 7T MR images, and the experimental results demonstrate that the joint use of entropy-based sampling strategy, vascular features and structured learning improves the segmentation accuracy, with the Dice similarity coefficient reaching 66%.

UR - http://www.scopus.com/inward/record.url?scp=84992491329&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84992491329&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-47157-0_8

DO - 10.1007/978-3-319-47157-0_8

M3 - Conference contribution

AN - SCOPUS:84992491329

SN - 9783319471563

VL - 10019 LNCS

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

SP - 61

EP - 68

BT - Machine Learning in Medical Imaging - 7th International Workshop, MLMI 2016 held in conjunction with MICCAI 2016, Proceedings

PB - Springer Verlag

ER -