Data-driven feature learning for myocardial segmentation of CP-BOLD MRI

Anirban Mukhopadhyay; Ilkay Oksuz; Marco Bevilacqua; Rohan Dharmakumar; Sotirios A. Tsaftaris

doi:10.1007/978-3-319-20309-6_22

Data-driven feature learning for myocardial segmentation of CP-BOLD MRI

Anirban Mukhopadhyay, Ilkay Oksuz^*, Marco Bevilacqua, Rohan Dharmakumar, Sotirios A. Tsaftaris

^*Corresponding author for this work

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

8 Citations (Scopus)

Abstract

Cardiac Phase-resolved Blood Oxygen-Level-Dependent (CP-BOLD) MR is capable of diagnosing an ongoing ischemia by detecting changes in myocardial intensity patterns at rest without any contrast and stress agents. Visualizing and detecting these changes require significant post-processing, including myocardial segmentation for isolating the myocardium. But, changes in myocardial intensity pattern and myocardial shape due to the heart’s motion challenge automated standard CINE MR myocardial segmentation techniques resulting in a significant drop of segmentation accuracy. We hypothesize that the main reason behind this phenomenon is the lack of discernible features. In this paper, a multi scale discriminative dictionary learning approach is proposed for supervised learning and sparse representation of the myocardium, to improve the myocardial feature selection. The technique is validated on a challenging dataset of CP-BOLD MR and standard CINE MR acquired in baseline and ischemic condition across 10 canine subjects. The proposed method significantly outperforms standard cardiac segmentation techniques, including segmentation via registration, level sets and supervised methods for myocardial segmentation.

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	189-197
Number of pages	9
ISBN (Print)	9783319203089, 9783319203089
DOIs	https://doi.org/10.1007/978-3-319-20309-6_22
Publication status	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

Bibliographical note

Publisher Copyright:
© Springer International Publishing Switzerland 2015.

Funding

This work was supported by the National Institutes of Health under Grant 2R01HL091989-05.

Funders	Funder number
National Institutes of Health	2R01HL091989-05

Keywords

CINE MR
CP-BOLD MR
Dictionary learning
Segmentation

Access to Document

10.1007/978-3-319-20309-6_22

Cite this

Mukhopadhyay, A., Oksuz, I., Bevilacqua, M., Dharmakumar, R., & Tsaftaris, S. A. (2015). Data-driven feature learning for myocardial segmentation of CP-BOLD MRI. 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. 189-197). (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_22

Mukhopadhyay, Anirban ; Oksuz, Ilkay ; Bevilacqua, Marco et al. / Data-driven feature learning for myocardial segmentation of CP-BOLD MRI. 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. 189-197 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{6de29d4807a14018a2d05ed7b7c23d44,

title = "Data-driven feature learning for myocardial segmentation of CP-BOLD MRI",

abstract = "Cardiac Phase-resolved Blood Oxygen-Level-Dependent (CP-BOLD) MR is capable of diagnosing an ongoing ischemia by detecting changes in myocardial intensity patterns at rest without any contrast and stress agents. Visualizing and detecting these changes require significant post-processing, including myocardial segmentation for isolating the myocardium. But, changes in myocardial intensity pattern and myocardial shape due to the heart{\textquoteright}s motion challenge automated standard CINE MR myocardial segmentation techniques resulting in a significant drop of segmentation accuracy. We hypothesize that the main reason behind this phenomenon is the lack of discernible features. In this paper, a multi scale discriminative dictionary learning approach is proposed for supervised learning and sparse representation of the myocardium, to improve the myocardial feature selection. The technique is validated on a challenging dataset of CP-BOLD MR and standard CINE MR acquired in baseline and ischemic condition across 10 canine subjects. The proposed method significantly outperforms standard cardiac segmentation techniques, including segmentation via registration, level sets and supervised methods for myocardial segmentation.",

keywords = "CINE MR, CP-BOLD MR, Dictionary learning, Segmentation",

author = "Anirban Mukhopadhyay and Ilkay Oksuz and Marco Bevilacqua and Rohan Dharmakumar and Tsaftaris, {Sotirios A.}",

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_22",

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 = "189--197",

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",

address = "Germany",

}

Mukhopadhyay, A, Oksuz, I, Bevilacqua, M, Dharmakumar, R & Tsaftaris, SA 2015, Data-driven feature learning for myocardial segmentation of CP-BOLD MRI. 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. 189-197, 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_22

Data-driven feature learning for myocardial segmentation of CP-BOLD MRI. / Mukhopadhyay, Anirban; Oksuz, Ilkay; Bevilacqua, Marco 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. 189-197 (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 - Data-driven feature learning for myocardial segmentation of CP-BOLD MRI

AU - Mukhopadhyay, Anirban

AU - Oksuz, Ilkay

AU - Bevilacqua, Marco

AU - Dharmakumar, Rohan

AU - Tsaftaris, Sotirios A.

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

PY - 2015

Y1 - 2015

N2 - Cardiac Phase-resolved Blood Oxygen-Level-Dependent (CP-BOLD) MR is capable of diagnosing an ongoing ischemia by detecting changes in myocardial intensity patterns at rest without any contrast and stress agents. Visualizing and detecting these changes require significant post-processing, including myocardial segmentation for isolating the myocardium. But, changes in myocardial intensity pattern and myocardial shape due to the heart’s motion challenge automated standard CINE MR myocardial segmentation techniques resulting in a significant drop of segmentation accuracy. We hypothesize that the main reason behind this phenomenon is the lack of discernible features. In this paper, a multi scale discriminative dictionary learning approach is proposed for supervised learning and sparse representation of the myocardium, to improve the myocardial feature selection. The technique is validated on a challenging dataset of CP-BOLD MR and standard CINE MR acquired in baseline and ischemic condition across 10 canine subjects. The proposed method significantly outperforms standard cardiac segmentation techniques, including segmentation via registration, level sets and supervised methods for myocardial segmentation.

AB - Cardiac Phase-resolved Blood Oxygen-Level-Dependent (CP-BOLD) MR is capable of diagnosing an ongoing ischemia by detecting changes in myocardial intensity patterns at rest without any contrast and stress agents. Visualizing and detecting these changes require significant post-processing, including myocardial segmentation for isolating the myocardium. But, changes in myocardial intensity pattern and myocardial shape due to the heart’s motion challenge automated standard CINE MR myocardial segmentation techniques resulting in a significant drop of segmentation accuracy. We hypothesize that the main reason behind this phenomenon is the lack of discernible features. In this paper, a multi scale discriminative dictionary learning approach is proposed for supervised learning and sparse representation of the myocardium, to improve the myocardial feature selection. The technique is validated on a challenging dataset of CP-BOLD MR and standard CINE MR acquired in baseline and ischemic condition across 10 canine subjects. The proposed method significantly outperforms standard cardiac segmentation techniques, including segmentation via registration, level sets and supervised methods for myocardial segmentation.

KW - CINE MR

KW - CP-BOLD MR

KW - Dictionary learning

KW - Segmentation

UR - http://www.scopus.com/inward/record.url?scp=84958550654&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-20309-6_22

DO - 10.1007/978-3-319-20309-6_22

M3 - Conference contribution

AN - SCOPUS:84958550654

SN - 9783319203089

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

SP - 189

EP - 197

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 -

Mukhopadhyay A, Oksuz I, Bevilacqua M, Dharmakumar R, Tsaftaris SA. Data-driven feature learning for myocardial segmentation of CP-BOLD MRI. 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. 189-197. (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_22

Data-driven feature learning for myocardial segmentation of CP-BOLD MRI

Abstract

Publication series

Conference

Bibliographical note

Funding

Keywords

Access to Document

Other files and links

Fingerprint

Cite this