Multi-modal Latent-Space Self-alignment for Super-Resolution Cardiac MR Segmentation

Yu Deng; Yang Wen; Linglong Qian; Esther Puyol Anton; Hao Xu; Kuberan Pushparajah; Zina Ibrahim; Richard Dobson; Alistair Young

doi:10.1007/978-3-031-23443-9_3

Multi-modal Latent-Space Self-alignment for Super-Resolution Cardiac MR Segmentation

Yu Deng^*
, Yang Wen
, Linglong Qian
, Esther Puyol Anton
, Hao Xu
, Kuberan Pushparajah
, Zina Ibrahim
, Richard Dobson
, Alistair Young

^*Corresponding author for this work

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

1 Citation (Scopus)

Abstract

2D cardiac MR cine images provide data with a high signal-to-noise ratio for the segmentation and reconstruction of the heart. These images are frequently used in clinical practice and research. However, the segments have low resolution in the through-plane direction, and standard interpolation methods are unable to improve resolution and precision. We proposed an end-to-end pipeline for producing high-resolution segments from 2D MR images. This pipeline utilised a bilateral optical flow warping method to recover images in the through-plane direction, while a SegResNet automatically generated segments of the left and right ventricles. A multi-modal latent-space self-alignment network was implemented to guarantee that the segments maintain an anatomical prior derived from unpaired 3D high-resolution CT scans. On 3D MR angiograms, the trained pipeline produced high-resolution segments that preserve an anatomical prior derived from patients with various cardiovascular diseases.

Original language	English
Title of host publication	Statistical Atlases and Computational Models of the Heart. Regular and CMRxMotion Challenge Papers - 13th International Workshop, STACOM 2022, Held in Conjunction with MICCAI 2022, Revised Selected Papers
Editors	Oscar Camara, Esther Puyol-Antón, Avan Suinesiaputra, Alistair Young, Chen Qin, Maxime Sermesant, Shuo Wang
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	26-35
Number of pages	10
ISBN (Print)	9783031234422
DOIs	https://doi.org/10.1007/978-3-031-23443-9_3
Publication status	Published - 2022
Externally published	Yes
Event	13th International Workshop on Statistical Atlases and Computational Models of the Heart, STACOM 2022, held in conjunction with the 25th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2022 - Singapore, Singapore Duration: 18 Sept 2022 → 18 Sept 2022

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	13593 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	13th International Workshop on Statistical Atlases and Computational Models of the Heart, STACOM 2022, held in conjunction with the 25th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2022
Country/Territory	Singapore
City	Singapore
Period	18/09/22 → 18/09/22

Keywords

CT angiogram
Cardiac MR
Domain adaptation
Super-resolution segmentation
VAE

Access to Document

10.1007/978-3-031-23443-9_3

Cite this

Deng, Y., Wen, Y., Qian, L., Puyol Anton, E., Xu, H., Pushparajah, K., Ibrahim, Z., Dobson, R., & Young, A. (2022). Multi-modal Latent-Space Self-alignment for Super-Resolution Cardiac MR Segmentation. In O. Camara, E. Puyol-Antón, A. Suinesiaputra, A. Young, C. Qin, M. Sermesant, & S. Wang (Eds.), Statistical Atlases and Computational Models of the Heart. Regular and CMRxMotion Challenge Papers - 13th International Workshop, STACOM 2022, Held in Conjunction with MICCAI 2022, Revised Selected Papers (pp. 26-35). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 13593 LNCS). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-031-23443-9_3

Deng, Yu ; Wen, Yang ; Qian, Linglong et al. / Multi-modal Latent-Space Self-alignment for Super-Resolution Cardiac MR Segmentation. Statistical Atlases and Computational Models of the Heart. Regular and CMRxMotion Challenge Papers - 13th International Workshop, STACOM 2022, Held in Conjunction with MICCAI 2022, Revised Selected Papers. editor / Oscar Camara ; Esther Puyol-Antón ; Avan Suinesiaputra ; Alistair Young ; Chen Qin ; Maxime Sermesant ; Shuo Wang. Springer Science and Business Media Deutschland GmbH, 2022. pp. 26-35 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{263884a3c90349dfb19abbe90b5eb312,

title = "Multi-modal Latent-Space Self-alignment for Super-Resolution Cardiac MR Segmentation",

abstract = "2D cardiac MR cine images provide data with a high signal-to-noise ratio for the segmentation and reconstruction of the heart. These images are frequently used in clinical practice and research. However, the segments have low resolution in the through-plane direction, and standard interpolation methods are unable to improve resolution and precision. We proposed an end-to-end pipeline for producing high-resolution segments from 2D MR images. This pipeline utilised a bilateral optical flow warping method to recover images in the through-plane direction, while a SegResNet automatically generated segments of the left and right ventricles. A multi-modal latent-space self-alignment network was implemented to guarantee that the segments maintain an anatomical prior derived from unpaired 3D high-resolution CT scans. On 3D MR angiograms, the trained pipeline produced high-resolution segments that preserve an anatomical prior derived from patients with various cardiovascular diseases.",

keywords = "CT angiogram, Cardiac MR, Domain adaptation, Super-resolution segmentation, VAE",

author = "Yu Deng and Yang Wen and Linglong Qian and \{Puyol Anton\}, Esther and Hao Xu and Kuberan Pushparajah and Zina Ibrahim and Richard Dobson and Alistair Young",

note = "Publisher Copyright: {\textcopyright} 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.; 13th International Workshop on Statistical Atlases and Computational Models of the Heart, STACOM 2022, held in conjunction with the 25th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2022 ; Conference date: 18-09-2022 Through 18-09-2022",

year = "2022",

doi = "10.1007/978-3-031-23443-9\_3",

language = "English",

isbn = "9783031234422",

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

publisher = "Springer Science and Business Media Deutschland GmbH",

pages = "26--35",

editor = "Oscar Camara and Esther Puyol-Ant{\'o}n and Avan Suinesiaputra and Alistair Young and Chen Qin and Maxime Sermesant and Shuo Wang",

booktitle = "Statistical Atlases and Computational Models of the Heart. Regular and CMRxMotion Challenge Papers - 13th International Workshop, STACOM 2022, Held in Conjunction with MICCAI 2022, Revised Selected Papers",

address = "Germany",

}

Deng, Y, Wen, Y, Qian, L, Puyol Anton, E, Xu, H, Pushparajah, K, Ibrahim, Z, Dobson, R & Young, A 2022, Multi-modal Latent-Space Self-alignment for Super-Resolution Cardiac MR Segmentation. in O Camara, E Puyol-Antón, A Suinesiaputra, A Young, C Qin, M Sermesant & S Wang (eds), Statistical Atlases and Computational Models of the Heart. Regular and CMRxMotion Challenge Papers - 13th International Workshop, STACOM 2022, Held in Conjunction with MICCAI 2022, Revised Selected Papers. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 13593 LNCS, Springer Science and Business Media Deutschland GmbH, pp. 26-35, 13th International Workshop on Statistical Atlases and Computational Models of the Heart, STACOM 2022, held in conjunction with the 25th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2022, Singapore, Singapore, 18/09/22. https://doi.org/10.1007/978-3-031-23443-9_3

Multi-modal Latent-Space Self-alignment for Super-Resolution Cardiac MR Segmentation. / Deng, Yu; Wen, Yang; Qian, Linglong et al.
Statistical Atlases and Computational Models of the Heart. Regular and CMRxMotion Challenge Papers - 13th International Workshop, STACOM 2022, Held in Conjunction with MICCAI 2022, Revised Selected Papers. ed. / Oscar Camara; Esther Puyol-Antón; Avan Suinesiaputra; Alistair Young; Chen Qin; Maxime Sermesant; Shuo Wang. Springer Science and Business Media Deutschland GmbH, 2022. p. 26-35 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 13593 LNCS).

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

TY - GEN

T1 - Multi-modal Latent-Space Self-alignment for Super-Resolution Cardiac MR Segmentation

AU - Deng, Yu

AU - Wen, Yang

AU - Qian, Linglong

AU - Puyol Anton, Esther

AU - Xu, Hao

AU - Pushparajah, Kuberan

AU - Ibrahim, Zina

AU - Dobson, Richard

AU - Young, Alistair

PY - 2022

Y1 - 2022

N2 - 2D cardiac MR cine images provide data with a high signal-to-noise ratio for the segmentation and reconstruction of the heart. These images are frequently used in clinical practice and research. However, the segments have low resolution in the through-plane direction, and standard interpolation methods are unable to improve resolution and precision. We proposed an end-to-end pipeline for producing high-resolution segments from 2D MR images. This pipeline utilised a bilateral optical flow warping method to recover images in the through-plane direction, while a SegResNet automatically generated segments of the left and right ventricles. A multi-modal latent-space self-alignment network was implemented to guarantee that the segments maintain an anatomical prior derived from unpaired 3D high-resolution CT scans. On 3D MR angiograms, the trained pipeline produced high-resolution segments that preserve an anatomical prior derived from patients with various cardiovascular diseases.

AB - 2D cardiac MR cine images provide data with a high signal-to-noise ratio for the segmentation and reconstruction of the heart. These images are frequently used in clinical practice and research. However, the segments have low resolution in the through-plane direction, and standard interpolation methods are unable to improve resolution and precision. We proposed an end-to-end pipeline for producing high-resolution segments from 2D MR images. This pipeline utilised a bilateral optical flow warping method to recover images in the through-plane direction, while a SegResNet automatically generated segments of the left and right ventricles. A multi-modal latent-space self-alignment network was implemented to guarantee that the segments maintain an anatomical prior derived from unpaired 3D high-resolution CT scans. On 3D MR angiograms, the trained pipeline produced high-resolution segments that preserve an anatomical prior derived from patients with various cardiovascular diseases.

KW - CT angiogram

KW - Cardiac MR

KW - Domain adaptation

KW - Super-resolution segmentation

KW - VAE

UR - https://www.scopus.com/pages/publications/85148024512

U2 - 10.1007/978-3-031-23443-9_3

DO - 10.1007/978-3-031-23443-9_3

M3 - Conference contribution

AN - SCOPUS:85148024512

SN - 9783031234422

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

SP - 26

EP - 35

BT - Statistical Atlases and Computational Models of the Heart. Regular and CMRxMotion Challenge Papers - 13th International Workshop, STACOM 2022, Held in Conjunction with MICCAI 2022, Revised Selected Papers

A2 - Camara, Oscar

A2 - Puyol-Antón, Esther

A2 - Suinesiaputra, Avan

A2 - Young, Alistair

A2 - Qin, Chen

A2 - Sermesant, Maxime

A2 - Wang, Shuo

PB - Springer Science and Business Media Deutschland GmbH

T2 - 13th International Workshop on Statistical Atlases and Computational Models of the Heart, STACOM 2022, held in conjunction with the 25th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2022

Y2 - 18 September 2022 through 18 September 2022

ER -

Deng Y, Wen Y, Qian L, Puyol Anton E, Xu H, Pushparajah K et al. Multi-modal Latent-Space Self-alignment for Super-Resolution Cardiac MR Segmentation. In Camara O, Puyol-Antón E, Suinesiaputra A, Young A, Qin C, Sermesant M, Wang S, editors, Statistical Atlases and Computational Models of the Heart. Regular and CMRxMotion Challenge Papers - 13th International Workshop, STACOM 2022, Held in Conjunction with MICCAI 2022, Revised Selected Papers. Springer Science and Business Media Deutschland GmbH. 2022. p. 26-35. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-031-23443-9_3