Standard
How Few Annotations are Needed for Segmentation Using a Multi-planar U-Net? / Laprade, William Michael; Perslev, Mathias; Sporring, Jon.
Deep Generative Models, and Data Augmentation, Labelling, and Imperfections - First Workshop, DGM4MICCAI 2021, and First Workshop, DALI 2021, Held in Conjunction with MICCAI 2021, Proceedings. red. / Sandy Engelhardt; Ilkay Oksuz; Dajiang Zhu; Yixuan Yuan; Anirban Mukhopadhyay; Nicholas Heller; Sharon Xiaolei Huang; Hien Nguyen; Raphael Sznitman; Yuan Xue. Springer, 2021. s. 209-216 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), Bind 13003 LNCS).
Publikation: Bidrag til bog/antologi/rapport › Konferencebidrag i proceedings › Forskning › fagfællebedømt
Harvard
Laprade, WM
, Perslev, M & Sporring, J 2021,
How Few Annotations are Needed for Segmentation Using a Multi-planar U-Net? i S Engelhardt, I Oksuz, D Zhu, Y Yuan, A Mukhopadhyay, N Heller, SX Huang, H Nguyen, R Sznitman & Y Xue (red),
Deep Generative Models, and Data Augmentation, Labelling, and Imperfections - First Workshop, DGM4MICCAI 2021, and First Workshop, DALI 2021, Held in Conjunction with MICCAI 2021, Proceedings. Springer, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), bind 13003 LNCS, s. 209-216, 1st Workshop on Deep Generative Models for Medical Image Computing and Computer Assisted Intervention, DGM4MICCAI 2021 and 1st Workshop on Data Augmentation, Labelling, and Imperfections, DALI 2021 held in conjunction with 24th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2021, Virtual, Online,
01/10/2021.
https://doi.org/10.1007/978-3-030-88210-5_20APA
Laprade, W. M.
, Perslev, M., & Sporring, J. (2021).
How Few Annotations are Needed for Segmentation Using a Multi-planar U-Net? I S. Engelhardt, I. Oksuz, D. Zhu, Y. Yuan, A. Mukhopadhyay, N. Heller, S. X. Huang, H. Nguyen, R. Sznitman, & Y. Xue (red.),
Deep Generative Models, and Data Augmentation, Labelling, and Imperfections - First Workshop, DGM4MICCAI 2021, and First Workshop, DALI 2021, Held in Conjunction with MICCAI 2021, Proceedings (s. 209-216). Springer. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) Bind 13003 LNCS
https://doi.org/10.1007/978-3-030-88210-5_20Vancouver
Laprade WM
, Perslev M, Sporring J.
How Few Annotations are Needed for Segmentation Using a Multi-planar U-Net? I Engelhardt S, Oksuz I, Zhu D, Yuan Y, Mukhopadhyay A, Heller N, Huang SX, Nguyen H, Sznitman R, Xue Y, red., Deep Generative Models, and Data Augmentation, Labelling, and Imperfections - First Workshop, DGM4MICCAI 2021, and First Workshop, DALI 2021, Held in Conjunction with MICCAI 2021, Proceedings. Springer. 2021. s. 209-216. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), Bind 13003 LNCS).
https://doi.org/10.1007/978-3-030-88210-5_20Author
Laprade, William Michael ; Perslev, Mathias ; Sporring, Jon. / How Few Annotations are Needed for Segmentation Using a Multi-planar U-Net?. Deep Generative Models, and Data Augmentation, Labelling, and Imperfections - First Workshop, DGM4MICCAI 2021, and First Workshop, DALI 2021, Held in Conjunction with MICCAI 2021, Proceedings. red. / Sandy Engelhardt ; Ilkay Oksuz ; Dajiang Zhu ; Yixuan Yuan ; Anirban Mukhopadhyay ; Nicholas Heller ; Sharon Xiaolei Huang ; Hien Nguyen ; Raphael Sznitman ; Yuan Xue. Springer, 2021. s. 209-216 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), Bind 13003 LNCS).
Bibtex
@inproceedings{79e60ba0cdd1445994c8dd5ec181a78f,
title = "How Few Annotations are Needed for Segmentation Using a Multi-planar U-Net?",
abstract = "U-Net architectures are an extremely powerful tool for segmenting 3D volumes, and the recently proposed multi-planar U-Net has reduced the computational requirement for using the U-Net architecture on three-dimensional isotropic data to a subset of two-dimensional planes. While multi-planar sampling considerably reduces the amount of training data needed, providing the required manually annotated data can still be a daunting task. In this article, we investigate the multi-planar U-Net{\textquoteright}s ability to learn three-dimensional structures in isotropic sampled images from sparsely annotated training samples. We extend the multi-planar U-Net with random annotations, and we present our empirical findings on two public domains, fully annotated by an expert. Surprisingly we find that the multi-planar U-Net on average outperforms the 3D U-Net in most cases in terms of dice, sensitivity, and specificity and that similar performance from the multi-planar unit can be obtained from half the number of annotations by doubling the number of automatically generated training planes. Thus, sometimes less is more!",
keywords = "3D imaging, Deep learning, Segmentation, Sparse annotations, U-Net",
author = "Laprade, {William Michael} and Mathias Perslev and Jon Sporring",
note = "Publisher Copyright: {\textcopyright} 2021, Springer Nature Switzerland AG.; 1st Workshop on Deep Generative Models for Medical Image Computing and Computer Assisted Intervention, DGM4MICCAI 2021 and 1st Workshop on Data Augmentation, Labelling, and Imperfections, DALI 2021 held in conjunction with 24th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2021 ; Conference date: 01-10-2021 Through 01-10-2021",
year = "2021",
doi = "10.1007/978-3-030-88210-5_20",
language = "English",
isbn = "9783030882099",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer",
pages = "209--216",
editor = "Sandy Engelhardt and Ilkay Oksuz and Dajiang Zhu and Yixuan Yuan and Anirban Mukhopadhyay and Nicholas Heller and Huang, {Sharon Xiaolei} and Hien Nguyen and Raphael Sznitman and Yuan Xue",
booktitle = "Deep Generative Models, and Data Augmentation, Labelling, and Imperfections - First Workshop, DGM4MICCAI 2021, and First Workshop, DALI 2021, Held in Conjunction with MICCAI 2021, Proceedings",
address = "Switzerland",
}
RIS
TY - GEN
T1 - How Few Annotations are Needed for Segmentation Using a Multi-planar U-Net?
AU - Laprade, William Michael
AU - Perslev, Mathias
AU - Sporring, Jon
N1 - Publisher Copyright:
© 2021, Springer Nature Switzerland AG.
PY - 2021
Y1 - 2021
N2 - U-Net architectures are an extremely powerful tool for segmenting 3D volumes, and the recently proposed multi-planar U-Net has reduced the computational requirement for using the U-Net architecture on three-dimensional isotropic data to a subset of two-dimensional planes. While multi-planar sampling considerably reduces the amount of training data needed, providing the required manually annotated data can still be a daunting task. In this article, we investigate the multi-planar U-Net’s ability to learn three-dimensional structures in isotropic sampled images from sparsely annotated training samples. We extend the multi-planar U-Net with random annotations, and we present our empirical findings on two public domains, fully annotated by an expert. Surprisingly we find that the multi-planar U-Net on average outperforms the 3D U-Net in most cases in terms of dice, sensitivity, and specificity and that similar performance from the multi-planar unit can be obtained from half the number of annotations by doubling the number of automatically generated training planes. Thus, sometimes less is more!
AB - U-Net architectures are an extremely powerful tool for segmenting 3D volumes, and the recently proposed multi-planar U-Net has reduced the computational requirement for using the U-Net architecture on three-dimensional isotropic data to a subset of two-dimensional planes. While multi-planar sampling considerably reduces the amount of training data needed, providing the required manually annotated data can still be a daunting task. In this article, we investigate the multi-planar U-Net’s ability to learn three-dimensional structures in isotropic sampled images from sparsely annotated training samples. We extend the multi-planar U-Net with random annotations, and we present our empirical findings on two public domains, fully annotated by an expert. Surprisingly we find that the multi-planar U-Net on average outperforms the 3D U-Net in most cases in terms of dice, sensitivity, and specificity and that similar performance from the multi-planar unit can be obtained from half the number of annotations by doubling the number of automatically generated training planes. Thus, sometimes less is more!
KW - 3D imaging
KW - Deep learning
KW - Segmentation
KW - Sparse annotations
KW - U-Net
U2 - 10.1007/978-3-030-88210-5_20
DO - 10.1007/978-3-030-88210-5_20
M3 - Article in proceedings
AN - SCOPUS:85116909289
SN - 9783030882099
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 209
EP - 216
BT - Deep Generative Models, and Data Augmentation, Labelling, and Imperfections - First Workshop, DGM4MICCAI 2021, and First Workshop, DALI 2021, Held in Conjunction with MICCAI 2021, Proceedings
A2 - Engelhardt, Sandy
A2 - Oksuz, Ilkay
A2 - Zhu, Dajiang
A2 - Yuan, Yixuan
A2 - Mukhopadhyay, Anirban
A2 - Heller, Nicholas
A2 - Huang, Sharon Xiaolei
A2 - Nguyen, Hien
A2 - Sznitman, Raphael
A2 - Xue, Yuan
PB - Springer
T2 - 1st Workshop on Deep Generative Models for Medical Image Computing and Computer Assisted Intervention, DGM4MICCAI 2021 and 1st Workshop on Data Augmentation, Labelling, and Imperfections, DALI 2021 held in conjunction with 24th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2021
Y2 - 1 October 2021 through 1 October 2021
ER -
