Difference between revisions of "Publications/xu.17.icip"

From LRDE

Line 28: Line 28:
 
neurological diseases in adults. Such studies are usually based on
 
neurological diseases in adults. Such studies are usually based on
 
quantitative analysis of different brain tissues, so it is essential
 
quantitative analysis of different brain tissues, so it is essential
to be able to classify accurately these tissues. In this paper, we
+
to be able to classify accurately these tissues. In this paper, we
propose a fast automatic method that segments 3D brain MR images
+
propose a fast automatic method that segments 3D brain MR images
into different tissues using fully convolutional network (FCN) and
+
into different tissues using fully convolutional network (FCN) and
transfer learning. As compared to existing deep learning-based
+
transfer learning. As compared to existing deep learning-based
approaches that rely either on 2D patches or on fully 3D FCN, our
+
approaches that rely either on 2D patches or on fully 3D FCN, our
method is way much faster: it only takes a few seconds, and only a
+
method is way much faster: it only takes a few seconds, and only a
single modality (T1 or T2) is required. In order to take the 3D
+
single modality (T1 or T2) is required. In order to take the 3D
information into account, the set of all 3 successive 2D slices are
+
information into account, the set of all 3 successive 2D slices are
stacked to form a set of 2D color images, which serve as input for
+
stacked to form a set of 2D color images, which serve as input for
the FCN pre-trained on ImageNet for natural image classification.
+
the FCN pre-trained on ImageNet for natural image classification.
To the best of our knowledge, this is the first method that attempts
+
To the best of our knowledge, this is the first method that attempts
to apply transfer learning for segmenting both neonatal and adult
+
to apply transfer learning for segmenting both neonatal and adult
brain 3D MR images, and we show that the proposed method achieves
+
brain 3D MR images, and we show that the proposed method achieves
state-of-the-art results on two public datasets.<nowiki>}</nowiki>,
+
state-of-the-art results on two public datasets.<nowiki>}</nowiki>,
 
note = <nowiki>{</nowiki>Submitted<nowiki>}</nowiki>,
 
note = <nowiki>{</nowiki>Submitted<nowiki>}</nowiki>,
 
optlrdepaper = <nowiki>{</nowiki>https://www.lrde.epita.fr/dload/papers/xu.2017.icip.pdf<nowiki>}</nowiki>
 
optlrdepaper = <nowiki>{</nowiki>https://www.lrde.epita.fr/dload/papers/xu.2017.icip.pdf<nowiki>}</nowiki>

Revision as of 14:44, 25 January 2017

Abstract

Brain magnetic resonance imaging (MRI) is widely used to assess brain developments in neonates and to diagnose a wide range of neurological diseases in adults. Such studies are usually based on quantitative analysis of different brain tissues, so it is essential to be able to classify accurately these tissues. In this paper, we propose a fast automatic method that segments 3D brain MR images into different tissues using fully convolutional network (FCN) and transfer learning. As compared to existing deep learning-based approaches that rely either on 2D patches or on fully 3D FCN, our method is way much faster: it only takes a few seconds, and only a single modality (T1 or T2) is required. In order to take the 3D information into account, the set of all 3 successive 2D slices are stacked to form a set of 2D color images, which serve as input for the FCN pre-trained on ImageNet for natural image classification. To the best of our knowledge, this is the first method that attempts to apply transfer learning for segmenting both neonatal and adult brain 3D MR images, and we show that the proposed method achieves state-of-the-art results on two public datasets.


Bibtex (lrde.bib)

@Article{	  xu.17.icip,
  author	= {Yongchao Xu and Thierry G\'eraud and Isabelle Bloch},
  title		= {From Neonatal to Adult Brain MR Image Segmentation in a
  Few Seconds Using 3D-Like Fully Convolutional Network and Transfer
  Learning},
  journal	= {Submitted for publication},
  year		= 2017,
  project	= {Image},
  abstract	= {Brain magnetic resonance imaging (MRI) is widely used to assess
                  brain developments in neonates and to diagnose a wide range of
                  neurological diseases in adults. Such studies are usually based on
                  quantitative analysis of different brain tissues, so it is essential
                  to be able to classify accurately these tissues.  In this paper, we
                  propose a fast automatic method that segments 3D brain MR images
                  into different tissues using fully convolutional network (FCN) and
                  transfer learning.  As compared to existing deep learning-based
                  approaches that rely either on 2D patches or on fully 3D FCN, our
                  method is way much faster: it only takes a few seconds, and only a
                  single modality (T1 or T2) is required.  In order to take the 3D
                  information into account, the set of all 3 successive 2D slices are
                  stacked to form a set of 2D color images, which serve as input for
                  the FCN pre-trained on ImageNet for natural image classification.
                  To the best of our knowledge, this is the first method that attempts
                  to apply transfer learning for segmenting both neonatal and adult
                  brain 3D MR images, and we show that the proposed method achieves
                  state-of-the-art results on two public datasets.},
  note		= {Submitted},
  optlrdepaper	= {https://www.lrde.epita.fr/dload/papers/xu.2017.icip.pdf}
		  
}