Deep convolutional neural networks for multi-modality isointense infant brain image segmentation

Wenlu Zhang, Rongjian Li, Houtao Deng, Li Wang, Weili Lin, Shuiwang Ji, Dinggang Shen

Research output: Contribution to journalArticle

294 Citations (Scopus)

Abstract

The segmentation of infant brain tissue images into white matter (WM), gray matter (GM), and cerebrospinal fluid (CSF) plays an important role in studying early brain development in health and disease. In the isointense stage (approximately 6-8 months of age), WM and GM exhibit similar levels of intensity in both T1 and T2 MR images, making the tissue segmentation very challenging. Only a small number of existing methods have been designed for tissue segmentation in this isointense stage; however, they only used a single T1 or T2 images, or the combination of T1 and T2 images. In this paper, we propose to use deep convolutional neural networks (CNNs) for segmenting isointense stage brain tissues using multi-modality MR images. CNNs are a type of deep models in which trainable filters and local neighborhood pooling operations are applied alternatingly on the raw input images, resulting in a hierarchy of increasingly complex features. Specifically, we used multi-modality information from T1, T2, and fractional anisotropy (FA) images as inputs and then generated the segmentation maps as outputs. The multiple intermediate layers applied convolution, pooling, normalization, and other operations to capture the highly nonlinear mappings between inputs and outputs. We compared the performance of our approach with that of the commonly used segmentation methods on a set of manually segmented isointense stage brain images. Results showed that our proposed model significantly outperformed prior methods on infant brain tissue segmentation. In addition, our results indicated that integration of multi-modality images led to significant performance improvement.

Original languageEnglish
Pages (from-to)214-224
Number of pages11
JournalNeuroImage
Volume108
DOIs
Publication statusPublished - 2015 Mar 1

Fingerprint

Brain
Anisotropy
Cerebrospinal Fluid
Health
White Matter
Gray Matter

Keywords

  • Convolutional neural networks
  • Deep learning
  • Image segmentation
  • Infant brain image
  • Multi-modality data

ASJC Scopus subject areas

  • Cognitive Neuroscience
  • Neurology

Cite this

Deep convolutional neural networks for multi-modality isointense infant brain image segmentation. / Zhang, Wenlu; Li, Rongjian; Deng, Houtao; Wang, Li; Lin, Weili; Ji, Shuiwang; Shen, Dinggang.

In: NeuroImage, Vol. 108, 01.03.2015, p. 214-224.

Research output: Contribution to journalArticle

Zhang, Wenlu ; Li, Rongjian ; Deng, Houtao ; Wang, Li ; Lin, Weili ; Ji, Shuiwang ; Shen, Dinggang. / Deep convolutional neural networks for multi-modality isointense infant brain image segmentation. In: NeuroImage. 2015 ; Vol. 108. pp. 214-224.
@article{32dc1cd286d54cc6a3f02ab3a36c115c,
title = "Deep convolutional neural networks for multi-modality isointense infant brain image segmentation",
abstract = "The segmentation of infant brain tissue images into white matter (WM), gray matter (GM), and cerebrospinal fluid (CSF) plays an important role in studying early brain development in health and disease. In the isointense stage (approximately 6-8 months of age), WM and GM exhibit similar levels of intensity in both T1 and T2 MR images, making the tissue segmentation very challenging. Only a small number of existing methods have been designed for tissue segmentation in this isointense stage; however, they only used a single T1 or T2 images, or the combination of T1 and T2 images. In this paper, we propose to use deep convolutional neural networks (CNNs) for segmenting isointense stage brain tissues using multi-modality MR images. CNNs are a type of deep models in which trainable filters and local neighborhood pooling operations are applied alternatingly on the raw input images, resulting in a hierarchy of increasingly complex features. Specifically, we used multi-modality information from T1, T2, and fractional anisotropy (FA) images as inputs and then generated the segmentation maps as outputs. The multiple intermediate layers applied convolution, pooling, normalization, and other operations to capture the highly nonlinear mappings between inputs and outputs. We compared the performance of our approach with that of the commonly used segmentation methods on a set of manually segmented isointense stage brain images. Results showed that our proposed model significantly outperformed prior methods on infant brain tissue segmentation. In addition, our results indicated that integration of multi-modality images led to significant performance improvement.",
keywords = "Convolutional neural networks, Deep learning, Image segmentation, Infant brain image, Multi-modality data",
author = "Wenlu Zhang and Rongjian Li and Houtao Deng and Li Wang and Weili Lin and Shuiwang Ji and Dinggang Shen",
year = "2015",
month = "3",
day = "1",
doi = "10.1016/j.neuroimage.2014.12.061",
language = "English",
volume = "108",
pages = "214--224",
journal = "NeuroImage",
issn = "1053-8119",
publisher = "Academic Press Inc.",

}

TY - JOUR

T1 - Deep convolutional neural networks for multi-modality isointense infant brain image segmentation

AU - Zhang, Wenlu

AU - Li, Rongjian

AU - Deng, Houtao

AU - Wang, Li

AU - Lin, Weili

AU - Ji, Shuiwang

AU - Shen, Dinggang

PY - 2015/3/1

Y1 - 2015/3/1

N2 - The segmentation of infant brain tissue images into white matter (WM), gray matter (GM), and cerebrospinal fluid (CSF) plays an important role in studying early brain development in health and disease. In the isointense stage (approximately 6-8 months of age), WM and GM exhibit similar levels of intensity in both T1 and T2 MR images, making the tissue segmentation very challenging. Only a small number of existing methods have been designed for tissue segmentation in this isointense stage; however, they only used a single T1 or T2 images, or the combination of T1 and T2 images. In this paper, we propose to use deep convolutional neural networks (CNNs) for segmenting isointense stage brain tissues using multi-modality MR images. CNNs are a type of deep models in which trainable filters and local neighborhood pooling operations are applied alternatingly on the raw input images, resulting in a hierarchy of increasingly complex features. Specifically, we used multi-modality information from T1, T2, and fractional anisotropy (FA) images as inputs and then generated the segmentation maps as outputs. The multiple intermediate layers applied convolution, pooling, normalization, and other operations to capture the highly nonlinear mappings between inputs and outputs. We compared the performance of our approach with that of the commonly used segmentation methods on a set of manually segmented isointense stage brain images. Results showed that our proposed model significantly outperformed prior methods on infant brain tissue segmentation. In addition, our results indicated that integration of multi-modality images led to significant performance improvement.

AB - The segmentation of infant brain tissue images into white matter (WM), gray matter (GM), and cerebrospinal fluid (CSF) plays an important role in studying early brain development in health and disease. In the isointense stage (approximately 6-8 months of age), WM and GM exhibit similar levels of intensity in both T1 and T2 MR images, making the tissue segmentation very challenging. Only a small number of existing methods have been designed for tissue segmentation in this isointense stage; however, they only used a single T1 or T2 images, or the combination of T1 and T2 images. In this paper, we propose to use deep convolutional neural networks (CNNs) for segmenting isointense stage brain tissues using multi-modality MR images. CNNs are a type of deep models in which trainable filters and local neighborhood pooling operations are applied alternatingly on the raw input images, resulting in a hierarchy of increasingly complex features. Specifically, we used multi-modality information from T1, T2, and fractional anisotropy (FA) images as inputs and then generated the segmentation maps as outputs. The multiple intermediate layers applied convolution, pooling, normalization, and other operations to capture the highly nonlinear mappings between inputs and outputs. We compared the performance of our approach with that of the commonly used segmentation methods on a set of manually segmented isointense stage brain images. Results showed that our proposed model significantly outperformed prior methods on infant brain tissue segmentation. In addition, our results indicated that integration of multi-modality images led to significant performance improvement.

KW - Convolutional neural networks

KW - Deep learning

KW - Image segmentation

KW - Infant brain image

KW - Multi-modality data

UR - http://www.scopus.com/inward/record.url?scp=84921492033&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84921492033&partnerID=8YFLogxK

U2 - 10.1016/j.neuroimage.2014.12.061

DO - 10.1016/j.neuroimage.2014.12.061

M3 - Article

C2 - 25562829

AN - SCOPUS:84921492033

VL - 108

SP - 214

EP - 224

JO - NeuroImage

JF - NeuroImage

SN - 1053-8119

ER -