Comparing deep learning strategies and attention mechanisms of discrete registration for multimodal image-guided interventions

In Young Ha; Mattias P. Heinrich

doi:10.1007/978-3-030-33642-4_16

Comparing deep learning strategies and attention mechanisms of discrete registration for multimodal image-guided interventions

In Young Ha^*, Mattias P. Heinrich

^*Corresponding author for this work

Institute of Medical Informatics

Abstract

In medical imaging, deep learning has been applied to segmentation and classification tasks successfully, whereas its use for image registration tasks is still limited. The use of discrete registration can alleviate the problems limiting the use of CNN based registration for large displacements by helping to capture more complex deformations. We evaluate different building blocks of learning based discrete registration for the CuRIOUS multimodal image registration challenge. We also propose a new attention module, which estimates information contents of a grid point, compare different loss functions and evaluate the influence of self-supervised pre-training of feature extraction step.

Original language	English
Title of host publication	LABELS 2019, HAL-MICCAI 2019, CuRIOUS 2019: Large-Scale Annotation of Biomedical Data and Expert Label Synthesis and Hardware Aware Learning for Medical Imaging and Computer Assisted Intervention
Editors	Luping Zhou, Nicholas Heller, Yiyu Shi, Yiming Xiao, Raphael Sznitman, Veronika Cheplygina, Diana Mateus, Emanuele Trucco, X. Sharon Hu, Danny Chen, Matthieu Chabanas, Hassan Rivaz, Ingerid Reinertsen
Number of pages	7
Volume	11851 LNCS
Publisher	Springer, Cham
Publication date	24.10.2019
Pages	145-151
ISBN (Print)	978-3-030-33641-7
ISBN (Electronic)	978-3-030-33642-4
DOIs	https://doi.org/10.1007/978-3-030-33642-4_16
Publication status	Published - 24.10.2019
Event	22nd International Conference on Medical Image Computing and Computer-Assisted Intervention - Shenzhen, China Duration: 13.10.2019 → 17.10.2019 Conference number: 232939

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

Access to Document

10.1007/978-3-030-33642-4_16

Cite this

Ha, I. Y., & Heinrich, M. P. (2019). Comparing deep learning strategies and attention mechanisms of discrete registration for multimodal image-guided interventions. In L. Zhou, N. Heller, Y. Shi, Y. Xiao, R. Sznitman, V. Cheplygina, D. Mateus, E. Trucco, X. S. Hu, D. Chen, M. Chabanas, H. Rivaz, & I. Reinertsen (Eds.), LABELS 2019, HAL-MICCAI 2019, CuRIOUS 2019: Large-Scale Annotation of Biomedical Data and Expert Label Synthesis and Hardware Aware Learning for Medical Imaging and Computer Assisted Intervention (Vol. 11851 LNCS, pp. 145-151). (Lecture Notes in Computer Science; Vol. 11851 LNCS). Springer, Cham. https://doi.org/10.1007/978-3-030-33642-4_16

Ha, In Young ; Heinrich, Mattias P. / Comparing deep learning strategies and attention mechanisms of discrete registration for multimodal image-guided interventions. LABELS 2019, HAL-MICCAI 2019, CuRIOUS 2019: Large-Scale Annotation of Biomedical Data and Expert Label Synthesis and Hardware Aware Learning for Medical Imaging and Computer Assisted Intervention. editor / Luping Zhou ; Nicholas Heller ; Yiyu Shi ; Yiming Xiao ; Raphael Sznitman ; Veronika Cheplygina ; Diana Mateus ; Emanuele Trucco ; X. Sharon Hu ; Danny Chen ; Matthieu Chabanas ; Hassan Rivaz ; Ingerid Reinertsen. Vol. 11851 LNCS Springer, Cham, 2019. pp. 145-151 (Lecture Notes in Computer Science).

@inproceedings{8bdda1aa065d4fcb914a801fa11979fc,

title = "Comparing deep learning strategies and attention mechanisms of discrete registration for multimodal image-guided interventions",

abstract = "In medical imaging, deep learning has been applied to segmentation and classification tasks successfully, whereas its use for image registration tasks is still limited. The use of discrete registration can alleviate the problems limiting the use of CNN based registration for large displacements by helping to capture more complex deformations. We evaluate different building blocks of learning based discrete registration for the CuRIOUS multimodal image registration challenge. We also propose a new attention module, which estimates information contents of a grid point, compare different loss functions and evaluate the influence of self-supervised pre-training of feature extraction step.",

author = "Ha, {In Young} and Heinrich, {Mattias P.}",

note = "Funding Information: This work is funded by the German Research Foundation DFG (HE 7364/2). Publisher Copyright: {\textcopyright} Springer Nature Switzerland AG 2019. Copyright: Copyright 2019 Elsevier B.V., All rights reserved.; 22nd International Conference on Medical Image Computing and Computer-Assisted Intervention , MICCAI 2019 ; Conference date: 13-10-2019 Through 17-10-2019",

year = "2019",

month = oct,

day = "24",

doi = "10.1007/978-3-030-33642-4_16",

language = "English",

isbn = "978-3-030-33641-7",

volume = "11851 LNCS",

series = "Lecture Notes in Computer Science",

publisher = "Springer, Cham",

pages = "145--151",

editor = "Zhou, {Luping } and Heller, {Nicholas } and Shi, {Yiyu } and Xiao, {Yiming } and Sznitman, {Raphael } and Cheplygina, {Veronika } and Mateus, {Diana } and Trucco, {Emanuele } and Hu, {X. Sharon } and Chen, {Danny } and Chabanas, {Matthieu } and Rivaz, {Hassan } and Reinertsen, {Ingerid }",

booktitle = "LABELS 2019, HAL-MICCAI 2019, CuRIOUS 2019: Large-Scale Annotation of Biomedical Data and Expert Label Synthesis and Hardware Aware Learning for Medical Imaging and Computer Assisted Intervention",

}

Ha, IY & Heinrich, MP 2019, Comparing deep learning strategies and attention mechanisms of discrete registration for multimodal image-guided interventions. in L Zhou, N Heller, Y Shi, Y Xiao, R Sznitman, V Cheplygina, D Mateus, E Trucco, XS Hu, D Chen, M Chabanas, H Rivaz & I Reinertsen (eds), LABELS 2019, HAL-MICCAI 2019, CuRIOUS 2019: Large-Scale Annotation of Biomedical Data and Expert Label Synthesis and Hardware Aware Learning for Medical Imaging and Computer Assisted Intervention. vol. 11851 LNCS, Lecture Notes in Computer Science, vol. 11851 LNCS, Springer, Cham, pp. 145-151, 22nd International Conference on Medical Image Computing and Computer-Assisted Intervention , Shenzhen, China, 13.10.19. https://doi.org/10.1007/978-3-030-33642-4_16

Comparing deep learning strategies and attention mechanisms of discrete registration for multimodal image-guided interventions. / Ha, In Young; Heinrich, Mattias P.
LABELS 2019, HAL-MICCAI 2019, CuRIOUS 2019: Large-Scale Annotation of Biomedical Data and Expert Label Synthesis and Hardware Aware Learning for Medical Imaging and Computer Assisted Intervention. ed. / Luping Zhou; Nicholas Heller; Yiyu Shi; Yiming Xiao; Raphael Sznitman; Veronika Cheplygina; Diana Mateus; Emanuele Trucco; X. Sharon Hu; Danny Chen; Matthieu Chabanas; Hassan Rivaz; Ingerid Reinertsen. Vol. 11851 LNCS Springer, Cham, 2019. p. 145-151 (Lecture Notes in Computer Science; Vol. 11851 LNCS).

Research output: Chapters in Books/Reports/Conference Proceedings › Conference contribution › peer-review

TY - GEN

T1 - Comparing deep learning strategies and attention mechanisms of discrete registration for multimodal image-guided interventions

AU - Ha, In Young

AU - Heinrich, Mattias P.

N1 - Conference code: 232939

PY - 2019/10/24

Y1 - 2019/10/24

N2 - In medical imaging, deep learning has been applied to segmentation and classification tasks successfully, whereas its use for image registration tasks is still limited. The use of discrete registration can alleviate the problems limiting the use of CNN based registration for large displacements by helping to capture more complex deformations. We evaluate different building blocks of learning based discrete registration for the CuRIOUS multimodal image registration challenge. We also propose a new attention module, which estimates information contents of a grid point, compare different loss functions and evaluate the influence of self-supervised pre-training of feature extraction step.

AB - In medical imaging, deep learning has been applied to segmentation and classification tasks successfully, whereas its use for image registration tasks is still limited. The use of discrete registration can alleviate the problems limiting the use of CNN based registration for large displacements by helping to capture more complex deformations. We evaluate different building blocks of learning based discrete registration for the CuRIOUS multimodal image registration challenge. We also propose a new attention module, which estimates information contents of a grid point, compare different loss functions and evaluate the influence of self-supervised pre-training of feature extraction step.

UR - https://www.scopus.com/record/display.uri?eid=2-s2.0-85076701068&origin=inward&txGid=af3bac79ec4742f2b7f9f22bb2a29998

U2 - 10.1007/978-3-030-33642-4_16

DO - 10.1007/978-3-030-33642-4_16

M3 - Conference contribution

SN - 978-3-030-33641-7

VL - 11851 LNCS

T3 - Lecture Notes in Computer Science

SP - 145

EP - 151

BT - LABELS 2019, HAL-MICCAI 2019, CuRIOUS 2019: Large-Scale Annotation of Biomedical Data and Expert Label Synthesis and Hardware Aware Learning for Medical Imaging and Computer Assisted Intervention

A2 - Zhou, Luping

A2 - Heller, Nicholas

A2 - Shi, Yiyu

A2 - Xiao, Yiming

A2 - Sznitman, Raphael

A2 - Cheplygina, Veronika

A2 - Mateus, Diana

A2 - Trucco, Emanuele

A2 - Hu, X. Sharon

A2 - Chen, Danny

A2 - Chabanas, Matthieu

A2 - Rivaz, Hassan

A2 - Reinertsen, Ingerid

PB - Springer, Cham

T2 - 22nd International Conference on Medical Image Computing and Computer-Assisted Intervention

Y2 - 13 October 2019 through 17 October 2019

ER -

Ha IY, Heinrich MP. Comparing deep learning strategies and attention mechanisms of discrete registration for multimodal image-guided interventions. In Zhou L, Heller N, Shi Y, Xiao Y, Sznitman R, Cheplygina V, Mateus D, Trucco E, Hu XS, Chen D, Chabanas M, Rivaz H, Reinertsen I, editors, LABELS 2019, HAL-MICCAI 2019, CuRIOUS 2019: Large-Scale Annotation of Biomedical Data and Expert Label Synthesis and Hardware Aware Learning for Medical Imaging and Computer Assisted Intervention. Vol. 11851 LNCS. Springer, Cham. 2019. p. 145-151. (Lecture Notes in Computer Science). doi: 10.1007/978-3-030-33642-4_16

Comparing deep learning strategies and attention mechanisms of discrete registration for multimodal image-guided interventions

Abstract

UN SDGs

Access to Document

Other files and links

Fingerprint

Cite this