Skip to main content

Lund University Publications

LUND UNIVERSITY LIBRARIES

Robust abdominal organ segmentation using regional convolutional neural networks

Larsson, Måns ; Zhang, Yuhang and Kahl, Fredrik LU (2017) 20th Scandinavian Conference on Image Analysis, SCIA 2017 In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) 10270 LNCS. p.41-52
Abstract

A fully automatic system for abdominal organ segmentation is presented. As a first step, an organ localization is obtained via a robust and efficient feature registration method where the center of the organ is estimated together with a region of interest surrounding the center. Then, a convolutional neural network performing voxelwise classification is applied. The convolutional neural network consists of several full 3D convolutional layers and takes both low and high resolution image data as input, which is designed to ensure both local and global consistency. Despite limited training data, our experimental results are on par with state-of-the-art approaches that have been developed over many years. More specifically the method is... (More)

A fully automatic system for abdominal organ segmentation is presented. As a first step, an organ localization is obtained via a robust and efficient feature registration method where the center of the organ is estimated together with a region of interest surrounding the center. Then, a convolutional neural network performing voxelwise classification is applied. The convolutional neural network consists of several full 3D convolutional layers and takes both low and high resolution image data as input, which is designed to ensure both local and global consistency. Despite limited training data, our experimental results are on par with state-of-the-art approaches that have been developed over many years. More specifically the method is applied to the MICCAI2015 challenge “Multi-Atlas Labeling Beyond the Cranial Vault” in the free competition for organ segmentation in the abdomen. It achieved the best results for 3 out of the 13 organs with a total mean Dice coefficient of 0.757 for all organs. Top scores were achieved for the gallbladder, the aorta and the right adrenal gland.

(Less)
Please use this url to cite or link to this publication:
author
; and
organization
publishing date
type
Chapter in Book/Report/Conference proceeding
publication status
published
subject
keywords
Convolutional neural networks, Medical image analysis, Segmentation
host publication
Image Analysis - 20th Scandinavian Conference, SCIA 2017, Proceedings
series title
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
volume
10270 LNCS
pages
12 pages
publisher
Springer
conference name
20th Scandinavian Conference on Image Analysis, SCIA 2017
conference location
Tromso, Norway
conference dates
2017-06-12 - 2017-06-14
external identifiers
  • scopus:85020380711
ISSN
03029743
16113349
ISBN
9783319591285
DOI
10.1007/978-3-319-59129-2_4
language
English
LU publication?
yes
id
aa51f526-4e38-459d-91b7-77ebdfcd95bb
date added to LUP
2017-06-30 09:03:19
date last changed
2024-01-28 21:10:39
@inproceedings{aa51f526-4e38-459d-91b7-77ebdfcd95bb,
  abstract     = {{<p>A fully automatic system for abdominal organ segmentation is presented. As a first step, an organ localization is obtained via a robust and efficient feature registration method where the center of the organ is estimated together with a region of interest surrounding the center. Then, a convolutional neural network performing voxelwise classification is applied. The convolutional neural network consists of several full 3D convolutional layers and takes both low and high resolution image data as input, which is designed to ensure both local and global consistency. Despite limited training data, our experimental results are on par with state-of-the-art approaches that have been developed over many years. More specifically the method is applied to the MICCAI2015 challenge “Multi-Atlas Labeling Beyond the Cranial Vault” in the free competition for organ segmentation in the abdomen. It achieved the best results for 3 out of the 13 organs with a total mean Dice coefficient of 0.757 for all organs. Top scores were achieved for the gallbladder, the aorta and the right adrenal gland.</p>}},
  author       = {{Larsson, Måns and Zhang, Yuhang and Kahl, Fredrik}},
  booktitle    = {{Image Analysis - 20th Scandinavian Conference, SCIA 2017, Proceedings}},
  isbn         = {{9783319591285}},
  issn         = {{03029743}},
  keywords     = {{Convolutional neural networks; Medical image analysis; Segmentation}},
  language     = {{eng}},
  pages        = {{41--52}},
  publisher    = {{Springer}},
  series       = {{Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)}},
  title        = {{Robust abdominal organ segmentation using regional convolutional neural networks}},
  url          = {{http://dx.doi.org/10.1007/978-3-319-59129-2_4}},
  doi          = {{10.1007/978-3-319-59129-2_4}},
  volume       = {{10270 LNCS}},
  year         = {{2017}},
}