Abstract
Objective: The goal of this paper is to automatically segment perivascular spaces (PVSs) in brain from high-resolution 7T magnetic resonance (MR) images. Methods: We propose a structured-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 two categories, i.e., PVS and background. In addition, we propose a novel entropy-based sampling strategy to extract informative samples in the background for training an explicit classification model. Since the vascular filters can extract various vascular features, even thin and low-contrast structures can be effectively extracted from noisy backgrounds. Moreover, continuous and smooth segmentation results can be obtained by utilizing patch-based structured labels. Results: The performance of our proposed method is evaluated on 19 subjects with 7T MR images, with the Dice similarity coefficient reaching 66%. Conclusion: The joint use of entropy-based sampling strategy, vascular features, and structured learning can improve the segmentation accuracy. Significance: Instead of manual annotation, our method provides an automatic way for PVS segmentation. Moreover, our method can be potentially used for other vascular structure segmentation because of its data-driven property.
| Original language | English |
|---|---|
| Article number | 7865910 |
| Pages (from-to) | 2803-2812 |
| Number of pages | 10 |
| Journal | IEEE Transactions on Biomedical Engineering |
| Volume | 64 |
| Issue number | 12 |
| DOIs | |
| Publication status | Published - 2017 Dec 1 |
Fingerprint
Keywords
- 7T magnetic resonance (MR) images
- Perivascular spaces (PVSs)
- segmentation
- structured random forest (SRF)
- vascular features
ASJC Scopus subject areas
- Biomedical Engineering
Cite this
Structured Learning for 3-D Perivascular Space Segmentation Using Vascular Features. / Zhang, Jun; Gao, Yaozong; Park, Sang Hyun; Zong, Xiaopeng; Lin, Weili; Shen, Dinggang.
In: IEEE Transactions on Biomedical Engineering, Vol. 64, No. 12, 7865910, 01.12.2017, p. 2803-2812.Research output: Contribution to journal › Article
}
TY - JOUR
T1 - Structured Learning for 3-D Perivascular Space Segmentation Using Vascular Features
AU - Zhang, Jun
AU - Gao, Yaozong
AU - Park, Sang Hyun
AU - Zong, Xiaopeng
AU - Lin, Weili
AU - Shen, Dinggang
PY - 2017/12/1
Y1 - 2017/12/1
N2 - Objective: The goal of this paper is to automatically segment perivascular spaces (PVSs) in brain from high-resolution 7T magnetic resonance (MR) images. Methods: We propose a structured-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 two categories, i.e., PVS and background. In addition, we propose a novel entropy-based sampling strategy to extract informative samples in the background for training an explicit classification model. Since the vascular filters can extract various vascular features, even thin and low-contrast structures can be effectively extracted from noisy backgrounds. Moreover, continuous and smooth segmentation results can be obtained by utilizing patch-based structured labels. Results: The performance of our proposed method is evaluated on 19 subjects with 7T MR images, with the Dice similarity coefficient reaching 66%. Conclusion: The joint use of entropy-based sampling strategy, vascular features, and structured learning can improve the segmentation accuracy. Significance: Instead of manual annotation, our method provides an automatic way for PVS segmentation. Moreover, our method can be potentially used for other vascular structure segmentation because of its data-driven property.
AB - Objective: The goal of this paper is to automatically segment perivascular spaces (PVSs) in brain from high-resolution 7T magnetic resonance (MR) images. Methods: We propose a structured-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 two categories, i.e., PVS and background. In addition, we propose a novel entropy-based sampling strategy to extract informative samples in the background for training an explicit classification model. Since the vascular filters can extract various vascular features, even thin and low-contrast structures can be effectively extracted from noisy backgrounds. Moreover, continuous and smooth segmentation results can be obtained by utilizing patch-based structured labels. Results: The performance of our proposed method is evaluated on 19 subjects with 7T MR images, with the Dice similarity coefficient reaching 66%. Conclusion: The joint use of entropy-based sampling strategy, vascular features, and structured learning can improve the segmentation accuracy. Significance: Instead of manual annotation, our method provides an automatic way for PVS segmentation. Moreover, our method can be potentially used for other vascular structure segmentation because of its data-driven property.
KW - 7T magnetic resonance (MR) images
KW - Perivascular spaces (PVSs)
KW - segmentation
KW - structured random forest (SRF)
KW - vascular features
UR - http://www.scopus.com/inward/record.url?scp=85040444235&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85040444235&partnerID=8YFLogxK
U2 - 10.1109/TBME.2016.2638918
DO - 10.1109/TBME.2016.2638918
M3 - Article
C2 - 28362579
AN - SCOPUS:85040444235
VL - 64
SP - 2803
EP - 2812
JO - IEEE Transactions on Biomedical Engineering
JF - IEEE Transactions on Biomedical Engineering
SN - 0018-9294
IS - 12
M1 - 7865910
ER -