Multi-channel multi-scale fully convolutional network for 3D perivascular spaces segmentation in 7T MR images

Chunfeng Lian, Jun Zhang, Mingxia Liu, Xiaopeng Zong, Sheng Che Hung, Weili Lin, Dinggang Shen

Research output: Contribution to journalArticle

20 Citations (Scopus)

Abstract

Accurate segmentation of perivascular spaces (PVSs) is an important step for quantitative study of PVS morphology. However, since PVSs are the thin tubular structures with relatively low contrast and also the number of PVSs is often large, it is challenging and time-consuming for manual delineation of PVSs. Although several automatic/semi-automatic methods, especially the traditional learning-based approaches, have been proposed for segmentation of 3D PVSs, their performance often depends on the hand-crafted image features, as well as sophisticated preprocessing operations prior to segmentation (e.g., specially defined regions-of-interest (ROIs)). In this paper, a novel fully convolutional neural network (FCN) with no requirement of any specified hand-crafted features and ROIs is proposed for efficient segmentation of PVSs. Particularly, the original T2-weighted 7T magnetic resonance (MR) images are first filtered via a non-local Haar-transform-based line singularity representation method to enhance the thin tubular structures. Both the original and enhanced MR images are used as multi-channel inputs to complementarily provide detailed image information and enhanced tubular structural information for the localization of PVSs. Multi-scale features are then automatically learned to characterize the spatial associations between PVSs and adjacent brain tissues. Finally, the produced PVS probability maps are recursively loaded into the network as an additional channel of inputs to provide the auxiliary contextual information for further refining the segmentation results. The proposed multi-channel multi-scale FCN has been evaluated on the 7T brain MR images scanned from 20 subjects. The experimental results show its superior performance compared with several state-of-the-art methods.

Original languageEnglish
Pages (from-to)106-117
Number of pages12
JournalMedical Image Analysis
Volume46
DOIs
Publication statusPublished - 2018 May 1

Fingerprint

Magnetic resonance
Magnetic Resonance Spectroscopy
Brain
Hand
Neural networks
Refining
Mathematical transformations
Learning
Tissue

Keywords

  • 7T MR images
  • Deep learning
  • Fully convolutional networks
  • Perivascular spaces
  • Segmentation

ASJC Scopus subject areas

  • Radiological and Ultrasound Technology
  • Radiology Nuclear Medicine and imaging
  • Computer Vision and Pattern Recognition
  • Health Informatics
  • Computer Graphics and Computer-Aided Design

Cite this

Multi-channel multi-scale fully convolutional network for 3D perivascular spaces segmentation in 7T MR images. / Lian, Chunfeng; Zhang, Jun; Liu, Mingxia; Zong, Xiaopeng; Hung, Sheng Che; Lin, Weili; Shen, Dinggang.

In: Medical Image Analysis, Vol. 46, 01.05.2018, p. 106-117.

Research output: Contribution to journalArticle

Lian, Chunfeng ; Zhang, Jun ; Liu, Mingxia ; Zong, Xiaopeng ; Hung, Sheng Che ; Lin, Weili ; Shen, Dinggang. / Multi-channel multi-scale fully convolutional network for 3D perivascular spaces segmentation in 7T MR images. In: Medical Image Analysis. 2018 ; Vol. 46. pp. 106-117.
@article{6e13661a5ee04223b0c94e713ac2c629,
title = "Multi-channel multi-scale fully convolutional network for 3D perivascular spaces segmentation in 7T MR images",
abstract = "Accurate segmentation of perivascular spaces (PVSs) is an important step for quantitative study of PVS morphology. However, since PVSs are the thin tubular structures with relatively low contrast and also the number of PVSs is often large, it is challenging and time-consuming for manual delineation of PVSs. Although several automatic/semi-automatic methods, especially the traditional learning-based approaches, have been proposed for segmentation of 3D PVSs, their performance often depends on the hand-crafted image features, as well as sophisticated preprocessing operations prior to segmentation (e.g., specially defined regions-of-interest (ROIs)). In this paper, a novel fully convolutional neural network (FCN) with no requirement of any specified hand-crafted features and ROIs is proposed for efficient segmentation of PVSs. Particularly, the original T2-weighted 7T magnetic resonance (MR) images are first filtered via a non-local Haar-transform-based line singularity representation method to enhance the thin tubular structures. Both the original and enhanced MR images are used as multi-channel inputs to complementarily provide detailed image information and enhanced tubular structural information for the localization of PVSs. Multi-scale features are then automatically learned to characterize the spatial associations between PVSs and adjacent brain tissues. Finally, the produced PVS probability maps are recursively loaded into the network as an additional channel of inputs to provide the auxiliary contextual information for further refining the segmentation results. The proposed multi-channel multi-scale FCN has been evaluated on the 7T brain MR images scanned from 20 subjects. The experimental results show its superior performance compared with several state-of-the-art methods.",
keywords = "7T MR images, Deep learning, Fully convolutional networks, Perivascular spaces, Segmentation",
author = "Chunfeng Lian and Jun Zhang and Mingxia Liu and Xiaopeng Zong and Hung, {Sheng Che} and Weili Lin and Dinggang Shen",
year = "2018",
month = "5",
day = "1",
doi = "10.1016/j.media.2018.02.009",
language = "English",
volume = "46",
pages = "106--117",
journal = "Medical Image Analysis",
issn = "1361-8415",
publisher = "Elsevier",

}

TY - JOUR

T1 - Multi-channel multi-scale fully convolutional network for 3D perivascular spaces segmentation in 7T MR images

AU - Lian, Chunfeng

AU - Zhang, Jun

AU - Liu, Mingxia

AU - Zong, Xiaopeng

AU - Hung, Sheng Che

AU - Lin, Weili

AU - Shen, Dinggang

PY - 2018/5/1

Y1 - 2018/5/1

N2 - Accurate segmentation of perivascular spaces (PVSs) is an important step for quantitative study of PVS morphology. However, since PVSs are the thin tubular structures with relatively low contrast and also the number of PVSs is often large, it is challenging and time-consuming for manual delineation of PVSs. Although several automatic/semi-automatic methods, especially the traditional learning-based approaches, have been proposed for segmentation of 3D PVSs, their performance often depends on the hand-crafted image features, as well as sophisticated preprocessing operations prior to segmentation (e.g., specially defined regions-of-interest (ROIs)). In this paper, a novel fully convolutional neural network (FCN) with no requirement of any specified hand-crafted features and ROIs is proposed for efficient segmentation of PVSs. Particularly, the original T2-weighted 7T magnetic resonance (MR) images are first filtered via a non-local Haar-transform-based line singularity representation method to enhance the thin tubular structures. Both the original and enhanced MR images are used as multi-channel inputs to complementarily provide detailed image information and enhanced tubular structural information for the localization of PVSs. Multi-scale features are then automatically learned to characterize the spatial associations between PVSs and adjacent brain tissues. Finally, the produced PVS probability maps are recursively loaded into the network as an additional channel of inputs to provide the auxiliary contextual information for further refining the segmentation results. The proposed multi-channel multi-scale FCN has been evaluated on the 7T brain MR images scanned from 20 subjects. The experimental results show its superior performance compared with several state-of-the-art methods.

AB - Accurate segmentation of perivascular spaces (PVSs) is an important step for quantitative study of PVS morphology. However, since PVSs are the thin tubular structures with relatively low contrast and also the number of PVSs is often large, it is challenging and time-consuming for manual delineation of PVSs. Although several automatic/semi-automatic methods, especially the traditional learning-based approaches, have been proposed for segmentation of 3D PVSs, their performance often depends on the hand-crafted image features, as well as sophisticated preprocessing operations prior to segmentation (e.g., specially defined regions-of-interest (ROIs)). In this paper, a novel fully convolutional neural network (FCN) with no requirement of any specified hand-crafted features and ROIs is proposed for efficient segmentation of PVSs. Particularly, the original T2-weighted 7T magnetic resonance (MR) images are first filtered via a non-local Haar-transform-based line singularity representation method to enhance the thin tubular structures. Both the original and enhanced MR images are used as multi-channel inputs to complementarily provide detailed image information and enhanced tubular structural information for the localization of PVSs. Multi-scale features are then automatically learned to characterize the spatial associations between PVSs and adjacent brain tissues. Finally, the produced PVS probability maps are recursively loaded into the network as an additional channel of inputs to provide the auxiliary contextual information for further refining the segmentation results. The proposed multi-channel multi-scale FCN has been evaluated on the 7T brain MR images scanned from 20 subjects. The experimental results show its superior performance compared with several state-of-the-art methods.

KW - 7T MR images

KW - Deep learning

KW - Fully convolutional networks

KW - Perivascular spaces

KW - Segmentation

UR - http://www.scopus.com/inward/record.url?scp=85042862129&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85042862129&partnerID=8YFLogxK

U2 - 10.1016/j.media.2018.02.009

DO - 10.1016/j.media.2018.02.009

M3 - Article

C2 - 29518675

AN - SCOPUS:85042862129

VL - 46

SP - 106

EP - 117

JO - Medical Image Analysis

JF - Medical Image Analysis

SN - 1361-8415

ER -