Deep fusion net for multi-atlas segmentation: Application to cardiac MR images

Heran Yang; Jian Sun; Huibin Li; Lisheng Wang; Zongben Xu

doi:10.1007/978-3-319-46723-8_60

Deep fusion net for multi-atlas segmentation: Application to cardiac MR images

Heran Yang
, Jian Sun
, Huibin Li
, Lisheng Wang
, Zongben Xu

School of Mathematics and Statistics

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

38 Scopus citations

Abstract

Atlas selection and label fusion are two major challenges in multi-atlas segmentation. In this paper,we propose a novel deep fusion net for better solving these challenges. Deep fusion net is a deep architecture by concatenating a feature extraction subnet and a non-local patch-based label fusion (NL-PLF) subnet in a single network. This network is trained end-to-end for automatically learning deep features achieving optimal performance in a NL-PLF framework. The learned deep features are further utilized in defining a similarity measure for atlas selection. Experimental results on Cardiac MR images for left ventricular segmentation demonstrate that our approach is effective both in atlas selection and multi-atlas label fusion,and achieves state of the art in performance.

Original language	English
Title of host publication	Medical Image Computing and Computer-Assisted Intervention - MICCAI 2016 - 19th International Conference, Proceedings
Editors	Gozde Unal, Sebastian Ourselin, Leo Joskowicz, Mert R. Sabuncu, William Wells
Publisher	Springer Verlag
Pages	521-528
Number of pages	8
ISBN (Print)	9783319467221
DOIs	https://doi.org/10.1007/978-3-319-46723-8_60
State	Published - 2016

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	9901 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Keywords

Atlas selection
Deep fusion net
End-to-end training
Feature learning
Left ventricular segmentation
Multi-atlas segmentation

Access to Document

10.1007/978-3-319-46723-8_60

Cite this

Yang, H., Sun, J., Li, H., Wang, L., & Xu, Z. (2016). Deep fusion net for multi-atlas segmentation: Application to cardiac MR images. In G. Unal, S. Ourselin, L. Joskowicz, M. R. Sabuncu, & W. Wells (Eds.), Medical Image Computing and Computer-Assisted Intervention - MICCAI 2016 - 19th International Conference, Proceedings (pp. 521-528). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9901 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-319-46723-8_60

Yang, Heran ; Sun, Jian ; Li, Huibin et al. / Deep fusion net for multi-atlas segmentation : Application to cardiac MR images. Medical Image Computing and Computer-Assisted Intervention - MICCAI 2016 - 19th International Conference, Proceedings. editor / Gozde Unal ; Sebastian Ourselin ; Leo Joskowicz ; Mert R. Sabuncu ; William Wells. Springer Verlag, 2016. pp. 521-528 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{4ca6f46495c74c5b8c922b297d9e37d5,

title = "Deep fusion net for multi-atlas segmentation: Application to cardiac MR images",

abstract = "Atlas selection and label fusion are two major challenges in multi-atlas segmentation. In this paper,we propose a novel deep fusion net for better solving these challenges. Deep fusion net is a deep architecture by concatenating a feature extraction subnet and a non-local patch-based label fusion (NL-PLF) subnet in a single network. This network is trained end-to-end for automatically learning deep features achieving optimal performance in a NL-PLF framework. The learned deep features are further utilized in defining a similarity measure for atlas selection. Experimental results on Cardiac MR images for left ventricular segmentation demonstrate that our approach is effective both in atlas selection and multi-atlas label fusion,and achieves state of the art in performance.",

keywords = "Atlas selection, Deep fusion net, End-to-end training, Feature learning, Left ventricular segmentation, Multi-atlas segmentation",

author = "Heran Yang and Jian Sun and Huibin Li and Lisheng Wang and Zongben Xu",

note = "Publisher Copyright: {\textcopyright} Springer International Publishing AG 2016.",

year = "2016",

doi = "10.1007/978-3-319-46723-8\_60",

language = "英语",

isbn = "9783319467221",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Verlag",

pages = "521--528",

editor = "Gozde Unal and Sebastian Ourselin and Leo Joskowicz and Sabuncu, \{Mert R.\} and William Wells",

booktitle = "Medical Image Computing and Computer-Assisted Intervention - MICCAI 2016 - 19th International Conference, Proceedings",

}

Yang, H, Sun, J , Li, H, Wang, L & Xu, Z 2016, Deep fusion net for multi-atlas segmentation: Application to cardiac MR images. in G Unal, S Ourselin, L Joskowicz, MR Sabuncu & W Wells (eds), Medical Image Computing and Computer-Assisted Intervention - MICCAI 2016 - 19th International Conference, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 9901 LNCS, Springer Verlag, pp. 521-528. https://doi.org/10.1007/978-3-319-46723-8_60

Deep fusion net for multi-atlas segmentation: Application to cardiac MR images. / Yang, Heran; Sun, Jian ; Li, Huibin et al.
Medical Image Computing and Computer-Assisted Intervention - MICCAI 2016 - 19th International Conference, Proceedings. ed. / Gozde Unal; Sebastian Ourselin; Leo Joskowicz; Mert R. Sabuncu; William Wells. Springer Verlag, 2016. p. 521-528 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9901 LNCS).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Deep fusion net for multi-atlas segmentation

T2 - Application to cardiac MR images

AU - Yang, Heran

AU - Sun, Jian

AU - Li, Huibin

AU - Wang, Lisheng

AU - Xu, Zongben

N1 - Publisher Copyright: © Springer International Publishing AG 2016.

PY - 2016

Y1 - 2016

N2 - Atlas selection and label fusion are two major challenges in multi-atlas segmentation. In this paper,we propose a novel deep fusion net for better solving these challenges. Deep fusion net is a deep architecture by concatenating a feature extraction subnet and a non-local patch-based label fusion (NL-PLF) subnet in a single network. This network is trained end-to-end for automatically learning deep features achieving optimal performance in a NL-PLF framework. The learned deep features are further utilized in defining a similarity measure for atlas selection. Experimental results on Cardiac MR images for left ventricular segmentation demonstrate that our approach is effective both in atlas selection and multi-atlas label fusion,and achieves state of the art in performance.

AB - Atlas selection and label fusion are two major challenges in multi-atlas segmentation. In this paper,we propose a novel deep fusion net for better solving these challenges. Deep fusion net is a deep architecture by concatenating a feature extraction subnet and a non-local patch-based label fusion (NL-PLF) subnet in a single network. This network is trained end-to-end for automatically learning deep features achieving optimal performance in a NL-PLF framework. The learned deep features are further utilized in defining a similarity measure for atlas selection. Experimental results on Cardiac MR images for left ventricular segmentation demonstrate that our approach is effective both in atlas selection and multi-atlas label fusion,and achieves state of the art in performance.

KW - Atlas selection

KW - Deep fusion net

KW - End-to-end training

KW - Feature learning

KW - Left ventricular segmentation

KW - Multi-atlas segmentation

UR - https://www.scopus.com/pages/publications/84996602227

U2 - 10.1007/978-3-319-46723-8_60

DO - 10.1007/978-3-319-46723-8_60

M3 - 会议稿件

AN - SCOPUS:84996602227

SN - 9783319467221

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 521

EP - 528

BT - Medical Image Computing and Computer-Assisted Intervention - MICCAI 2016 - 19th International Conference, Proceedings

A2 - Unal, Gozde

A2 - Ourselin, Sebastian

A2 - Joskowicz, Leo

A2 - Sabuncu, Mert R.

A2 - Wells, William

PB - Springer Verlag

ER -

Yang H, Sun J , Li H, Wang L, Xu Z. Deep fusion net for multi-atlas segmentation: Application to cardiac MR images. In Unal G, Ourselin S, Joskowicz L, Sabuncu MR, Wells W, editors, Medical Image Computing and Computer-Assisted Intervention - MICCAI 2016 - 19th International Conference, Proceedings. Springer Verlag. 2016. p. 521-528. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-319-46723-8_60

Deep fusion net for multi-atlas segmentation: Application to cardiac MR images

Abstract

Publication series

Keywords

Access to Document

Other files and links

Fingerprint

Cite this