Segmentation of Organs at Risk in thoracic CT images using a SharpMask architecture and Conditional Random Fields

R. Trullo, C. Petitjean, S. Ruan, B. Dubray, D. Nie, Dinggang Shen

Research output: Chapter in Book/Report/Conference proceedingConference contribution

23 Citations (Scopus)

Abstract

Cancer is one of the leading causes of death worldwide. Radiotherapy is a standard treatment for this condition and the first step of the radiotherapy process is to identify the target volumes to be targeted and the healthy organs at risk (OAR) to be protected. Unlike previous methods for automatic segmentation of OAR that typically use local information and individually segment each OAR, in this paper, we propose a deep learning framework for the joint segmentation of OAR in CT images of the thorax, specifically the heart, esophagus, trachea and the aorta. Making use of Fully Convolutional Networks (FCN), we present several extensions that improve the performance, including a new architecture that allows to use low level features with high level information, effectively combining local and global information for improving the localization accuracy. Finally, by using Conditional Random Fields (specifically the CRF as Recurrent Neural Network model), we are able to account for relationships between the organs to further improve the segmentation results. Experiments demonstrate competitive performance on a dataset of 30 CT scans.

Original languageEnglish
Title of host publication2017 IEEE 14th International Symposium on Biomedical Imaging, ISBI 2017
PublisherIEEE Computer Society
Pages1003-1006
Number of pages4
ISBN (Electronic)9781509011711
DOIs
Publication statusPublished - 2017 Jun 15
Externally publishedYes
Event14th IEEE International Symposium on Biomedical Imaging, ISBI 2017 - Melbourne, Australia
Duration: 2017 Apr 182017 Apr 21

Other

Other14th IEEE International Symposium on Biomedical Imaging, ISBI 2017
CountryAustralia
CityMelbourne
Period17/4/1817/4/21

Fingerprint

Organs at Risk
Thorax
Radiotherapy
Neural Networks (Computer)
Computerized tomography
Recurrent neural networks
Trachea
Esophagus
Aorta
Cause of Death
Joints
Learning
Neoplasms
Experiments

Keywords

  • CRF
  • CRFasRNN
  • CT Segmentation
  • Fully Convolutional Networks (FCN)

ASJC Scopus subject areas

  • Biomedical Engineering
  • Radiology Nuclear Medicine and imaging

Cite this

Trullo, R., Petitjean, C., Ruan, S., Dubray, B., Nie, D., & Shen, D. (2017). Segmentation of Organs at Risk in thoracic CT images using a SharpMask architecture and Conditional Random Fields. In 2017 IEEE 14th International Symposium on Biomedical Imaging, ISBI 2017 (pp. 1003-1006). [7950685] IEEE Computer Society. https://doi.org/10.1109/ISBI.2017.7950685

Segmentation of Organs at Risk in thoracic CT images using a SharpMask architecture and Conditional Random Fields. / Trullo, R.; Petitjean, C.; Ruan, S.; Dubray, B.; Nie, D.; Shen, Dinggang.

2017 IEEE 14th International Symposium on Biomedical Imaging, ISBI 2017. IEEE Computer Society, 2017. p. 1003-1006 7950685.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Trullo, R, Petitjean, C, Ruan, S, Dubray, B, Nie, D & Shen, D 2017, Segmentation of Organs at Risk in thoracic CT images using a SharpMask architecture and Conditional Random Fields. in 2017 IEEE 14th International Symposium on Biomedical Imaging, ISBI 2017., 7950685, IEEE Computer Society, pp. 1003-1006, 14th IEEE International Symposium on Biomedical Imaging, ISBI 2017, Melbourne, Australia, 17/4/18. https://doi.org/10.1109/ISBI.2017.7950685
Trullo R, Petitjean C, Ruan S, Dubray B, Nie D, Shen D. Segmentation of Organs at Risk in thoracic CT images using a SharpMask architecture and Conditional Random Fields. In 2017 IEEE 14th International Symposium on Biomedical Imaging, ISBI 2017. IEEE Computer Society. 2017. p. 1003-1006. 7950685 https://doi.org/10.1109/ISBI.2017.7950685
Trullo, R. ; Petitjean, C. ; Ruan, S. ; Dubray, B. ; Nie, D. ; Shen, Dinggang. / Segmentation of Organs at Risk in thoracic CT images using a SharpMask architecture and Conditional Random Fields. 2017 IEEE 14th International Symposium on Biomedical Imaging, ISBI 2017. IEEE Computer Society, 2017. pp. 1003-1006
@inproceedings{219408e822bf4aeaba85f29546fd58ea,
title = "Segmentation of Organs at Risk in thoracic CT images using a SharpMask architecture and Conditional Random Fields",
abstract = "Cancer is one of the leading causes of death worldwide. Radiotherapy is a standard treatment for this condition and the first step of the radiotherapy process is to identify the target volumes to be targeted and the healthy organs at risk (OAR) to be protected. Unlike previous methods for automatic segmentation of OAR that typically use local information and individually segment each OAR, in this paper, we propose a deep learning framework for the joint segmentation of OAR in CT images of the thorax, specifically the heart, esophagus, trachea and the aorta. Making use of Fully Convolutional Networks (FCN), we present several extensions that improve the performance, including a new architecture that allows to use low level features with high level information, effectively combining local and global information for improving the localization accuracy. Finally, by using Conditional Random Fields (specifically the CRF as Recurrent Neural Network model), we are able to account for relationships between the organs to further improve the segmentation results. Experiments demonstrate competitive performance on a dataset of 30 CT scans.",
keywords = "CRF, CRFasRNN, CT Segmentation, Fully Convolutional Networks (FCN)",
author = "R. Trullo and C. Petitjean and S. Ruan and B. Dubray and D. Nie and Dinggang Shen",
year = "2017",
month = "6",
day = "15",
doi = "10.1109/ISBI.2017.7950685",
language = "English",
pages = "1003--1006",
booktitle = "2017 IEEE 14th International Symposium on Biomedical Imaging, ISBI 2017",
publisher = "IEEE Computer Society",

}

TY - GEN

T1 - Segmentation of Organs at Risk in thoracic CT images using a SharpMask architecture and Conditional Random Fields

AU - Trullo, R.

AU - Petitjean, C.

AU - Ruan, S.

AU - Dubray, B.

AU - Nie, D.

AU - Shen, Dinggang

PY - 2017/6/15

Y1 - 2017/6/15

N2 - Cancer is one of the leading causes of death worldwide. Radiotherapy is a standard treatment for this condition and the first step of the radiotherapy process is to identify the target volumes to be targeted and the healthy organs at risk (OAR) to be protected. Unlike previous methods for automatic segmentation of OAR that typically use local information and individually segment each OAR, in this paper, we propose a deep learning framework for the joint segmentation of OAR in CT images of the thorax, specifically the heart, esophagus, trachea and the aorta. Making use of Fully Convolutional Networks (FCN), we present several extensions that improve the performance, including a new architecture that allows to use low level features with high level information, effectively combining local and global information for improving the localization accuracy. Finally, by using Conditional Random Fields (specifically the CRF as Recurrent Neural Network model), we are able to account for relationships between the organs to further improve the segmentation results. Experiments demonstrate competitive performance on a dataset of 30 CT scans.

AB - Cancer is one of the leading causes of death worldwide. Radiotherapy is a standard treatment for this condition and the first step of the radiotherapy process is to identify the target volumes to be targeted and the healthy organs at risk (OAR) to be protected. Unlike previous methods for automatic segmentation of OAR that typically use local information and individually segment each OAR, in this paper, we propose a deep learning framework for the joint segmentation of OAR in CT images of the thorax, specifically the heart, esophagus, trachea and the aorta. Making use of Fully Convolutional Networks (FCN), we present several extensions that improve the performance, including a new architecture that allows to use low level features with high level information, effectively combining local and global information for improving the localization accuracy. Finally, by using Conditional Random Fields (specifically the CRF as Recurrent Neural Network model), we are able to account for relationships between the organs to further improve the segmentation results. Experiments demonstrate competitive performance on a dataset of 30 CT scans.

KW - CRF

KW - CRFasRNN

KW - CT Segmentation

KW - Fully Convolutional Networks (FCN)

UR - http://www.scopus.com/inward/record.url?scp=85023200536&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85023200536&partnerID=8YFLogxK

U2 - 10.1109/ISBI.2017.7950685

DO - 10.1109/ISBI.2017.7950685

M3 - Conference contribution

AN - SCOPUS:85023200536

SP - 1003

EP - 1006

BT - 2017 IEEE 14th International Symposium on Biomedical Imaging, ISBI 2017

PB - IEEE Computer Society

ER -