Probabilistic edge map (PEM) for 3D ultrasound image registration and multi-atlas left ventricle segmentation

Ozan Oktay; Alberto Gomez; Kevin Keraudren; Andreas Schuh; Wenjia Bai; Wenzhe Shi; Graeme Penney; Daniel Rueckert

doi:10.1007/978-3-319-20309-6_26

Probabilistic edge map (PEM) for 3D ultrasound image registration and multi-atlas left ventricle segmentation

Ozan Oktay, Alberto Gomez, Kevin Keraudren, Andreas Schuh, Wenjia Bai, Wenzhe Shi, Graeme Penney, Daniel Rueckert

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

6 Scopus citations

Abstract

Automated left ventricle (LV) segmentation in 3D ultrasound (3D-US) remains a challenging research problem due to variable image quality and limited field-of-view. Modern segmentation approaches (shape, appearance and contour model based surface fitting) require an accurate initialization and good image boundary features to obtain reliable and consistent results. They are therefore not well suited for this problem. The proposed method overcomes those limitations with a novel and generic 3D-US image boundary representation technique: Probabilistic Edge Map (PEM). This new representation captures regularized and complete edge responses from standard 3D-US images. PEM is utilized in a multi-atlas LV segmentation framework to spatially align target and atlas images. Experiments on data from the MICCAI CETUS challenge show that the proposed approach is better suited for LV segmentation than the active contour, appearance and voxel classification approaches, achieving lower surface distance errors and better LV volume estimates.

Original language	English
Title of host publication	Functional Imaging and Modeling of the Heart - 8th International Conference, FIMH 2015, Proceedings
Editors	Hans van Assen, Peter Bovendeerd, Hans van Assen, Peter Bovendeerd, Tammo Delhaas, Tammo Delhaas
Publisher	Springer Verlag
Pages	223-230
Number of pages	8
ISBN (Print)	9783319203089, 9783319203089
DOIs	https://doi.org/10.1007/978-3-319-20309-6_26
State	Published - 2015
Externally published	Yes
Event	8th International Conference on Functional Imaging and Modeling of the Heart, FIMH 2015 - Maastricht, Netherlands Duration: 25 Jun 2015 → 27 Jun 2015

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	9126
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	8th International Conference on Functional Imaging and Modeling of the Heart, FIMH 2015
Country/Territory	Netherlands
City	Maastricht
Period	25/06/15 → 27/06/15

Keywords

Left ventricle segmentation
Multiatlas label fusion
Probabilistic edge map
Structured decision forest
Ultrasound image analysis’

Access to Document

10.1007/978-3-319-20309-6_26

Cite this

Oktay, O., Gomez, A., Keraudren, K., Schuh, A., Bai, W., Shi, W., Penney, G., & Rueckert, D. (2015). Probabilistic edge map (PEM) for 3D ultrasound image registration and multi-atlas left ventricle segmentation. In H. van Assen, P. Bovendeerd, H. van Assen, P. Bovendeerd, T. Delhaas, & T. Delhaas (Eds.), Functional Imaging and Modeling of the Heart - 8th International Conference, FIMH 2015, Proceedings (pp. 223-230). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9126). Springer Verlag. https://doi.org/10.1007/978-3-319-20309-6_26

Oktay, Ozan ; Gomez, Alberto ; Keraudren, Kevin et al. / Probabilistic edge map (PEM) for 3D ultrasound image registration and multi-atlas left ventricle segmentation. Functional Imaging and Modeling of the Heart - 8th International Conference, FIMH 2015, Proceedings. editor / Hans van Assen ; Peter Bovendeerd ; Hans van Assen ; Peter Bovendeerd ; Tammo Delhaas ; Tammo Delhaas. Springer Verlag, 2015. pp. 223-230 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{24ab3fb12cce4b8a99b3e4edb27eda9c,

title = "Probabilistic edge map (PEM) for 3D ultrasound image registration and multi-atlas left ventricle segmentation",

abstract = "Automated left ventricle (LV) segmentation in 3D ultrasound (3D-US) remains a challenging research problem due to variable image quality and limited field-of-view. Modern segmentation approaches (shape, appearance and contour model based surface fitting) require an accurate initialization and good image boundary features to obtain reliable and consistent results. They are therefore not well suited for this problem. The proposed method overcomes those limitations with a novel and generic 3D-US image boundary representation technique: Probabilistic Edge Map (PEM). This new representation captures regularized and complete edge responses from standard 3D-US images. PEM is utilized in a multi-atlas LV segmentation framework to spatially align target and atlas images. Experiments on data from the MICCAI CETUS challenge show that the proposed approach is better suited for LV segmentation than the active contour, appearance and voxel classification approaches, achieving lower surface distance errors and better LV volume estimates.",

keywords = "Left ventricle segmentation, Multiatlas label fusion, Probabilistic edge map, Structured decision forest, Ultrasound image analysis{\textquoteright}",

author = "Ozan Oktay and Alberto Gomez and Kevin Keraudren and Andreas Schuh and Wenjia Bai and Wenzhe Shi and Graeme Penney and Daniel Rueckert",

note = "Publisher Copyright: {\textcopyright} Springer International Publishing Switzerland 2015.; 8th International Conference on Functional Imaging and Modeling of the Heart, FIMH 2015 ; Conference date: 25-06-2015 Through 27-06-2015",

year = "2015",

doi = "10.1007/978-3-319-20309-6_26",

language = "English",

isbn = "9783319203089",

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

publisher = "Springer Verlag",

pages = "223--230",

editor = "{van Assen}, Hans and Peter Bovendeerd and {van Assen}, Hans and Peter Bovendeerd and Tammo Delhaas and Tammo Delhaas",

booktitle = "Functional Imaging and Modeling of the Heart - 8th International Conference, FIMH 2015, Proceedings",

}

Oktay, O, Gomez, A, Keraudren, K, Schuh, A, Bai, W, Shi, W, Penney, G & Rueckert, D 2015, Probabilistic edge map (PEM) for 3D ultrasound image registration and multi-atlas left ventricle segmentation. in H van Assen, P Bovendeerd, H van Assen, P Bovendeerd, T Delhaas & T Delhaas (eds), Functional Imaging and Modeling of the Heart - 8th International Conference, FIMH 2015, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 9126, Springer Verlag, pp. 223-230, 8th International Conference on Functional Imaging and Modeling of the Heart, FIMH 2015, Maastricht, Netherlands, 25/06/15. https://doi.org/10.1007/978-3-319-20309-6_26

Probabilistic edge map (PEM) for 3D ultrasound image registration and multi-atlas left ventricle segmentation. / Oktay, Ozan; Gomez, Alberto; Keraudren, Kevin et al.
Functional Imaging and Modeling of the Heart - 8th International Conference, FIMH 2015, Proceedings. ed. / Hans van Assen; Peter Bovendeerd; Hans van Assen; Peter Bovendeerd; Tammo Delhaas; Tammo Delhaas. Springer Verlag, 2015. p. 223-230 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9126).

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

TY - GEN

T1 - Probabilistic edge map (PEM) for 3D ultrasound image registration and multi-atlas left ventricle segmentation

AU - Oktay, Ozan

AU - Gomez, Alberto

AU - Keraudren, Kevin

AU - Schuh, Andreas

AU - Bai, Wenjia

AU - Shi, Wenzhe

AU - Penney, Graeme

AU - Rueckert, Daniel

N1 - Publisher Copyright: © Springer International Publishing Switzerland 2015.

PY - 2015

Y1 - 2015

N2 - Automated left ventricle (LV) segmentation in 3D ultrasound (3D-US) remains a challenging research problem due to variable image quality and limited field-of-view. Modern segmentation approaches (shape, appearance and contour model based surface fitting) require an accurate initialization and good image boundary features to obtain reliable and consistent results. They are therefore not well suited for this problem. The proposed method overcomes those limitations with a novel and generic 3D-US image boundary representation technique: Probabilistic Edge Map (PEM). This new representation captures regularized and complete edge responses from standard 3D-US images. PEM is utilized in a multi-atlas LV segmentation framework to spatially align target and atlas images. Experiments on data from the MICCAI CETUS challenge show that the proposed approach is better suited for LV segmentation than the active contour, appearance and voxel classification approaches, achieving lower surface distance errors and better LV volume estimates.

AB - Automated left ventricle (LV) segmentation in 3D ultrasound (3D-US) remains a challenging research problem due to variable image quality and limited field-of-view. Modern segmentation approaches (shape, appearance and contour model based surface fitting) require an accurate initialization and good image boundary features to obtain reliable and consistent results. They are therefore not well suited for this problem. The proposed method overcomes those limitations with a novel and generic 3D-US image boundary representation technique: Probabilistic Edge Map (PEM). This new representation captures regularized and complete edge responses from standard 3D-US images. PEM is utilized in a multi-atlas LV segmentation framework to spatially align target and atlas images. Experiments on data from the MICCAI CETUS challenge show that the proposed approach is better suited for LV segmentation than the active contour, appearance and voxel classification approaches, achieving lower surface distance errors and better LV volume estimates.

KW - Left ventricle segmentation

KW - Multiatlas label fusion

KW - Probabilistic edge map

KW - Structured decision forest

KW - Ultrasound image analysis’

UR - http://www.scopus.com/inward/record.url?scp=84958520317&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-20309-6_26

DO - 10.1007/978-3-319-20309-6_26

M3 - Conference contribution

AN - SCOPUS:84958520317

SN - 9783319203089

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

SP - 223

EP - 230

BT - Functional Imaging and Modeling of the Heart - 8th International Conference, FIMH 2015, Proceedings

A2 - van Assen, Hans

A2 - Bovendeerd, Peter

A2 - van Assen, Hans

A2 - Bovendeerd, Peter

A2 - Delhaas, Tammo

PB - Springer Verlag

T2 - 8th International Conference on Functional Imaging and Modeling of the Heart, FIMH 2015

Y2 - 25 June 2015 through 27 June 2015

ER -

Oktay O, Gomez A, Keraudren K, Schuh A, Bai W, Shi W et al. Probabilistic edge map (PEM) for 3D ultrasound image registration and multi-atlas left ventricle segmentation. In van Assen H, Bovendeerd P, van Assen H, Bovendeerd P, Delhaas T, Delhaas T, editors, Functional Imaging and Modeling of the Heart - 8th International Conference, FIMH 2015, Proceedings. Springer Verlag. 2015. p. 223-230. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-319-20309-6_26

Probabilistic edge map (PEM) for 3D ultrasound image registration and multi-atlas left ventricle segmentation

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this