How Few Annotations are Needed for Segmentation Using a Multi-planar U-Net?
Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review
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!
Original language | English |
---|---|
Title of host publication | 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 |
Editors | Sandy Engelhardt, Ilkay Oksuz, Dajiang Zhu, Yixuan Yuan, Anirban Mukhopadhyay, Nicholas Heller, Sharon Xiaolei Huang, Hien Nguyen, Raphael Sznitman, Yuan Xue |
Publisher | Springer |
Publication date | 2021 |
Pages | 209-216 |
ISBN (Print) | 9783030882099 |
DOIs | |
Publication status | Published - 2021 |
Event | 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 Duration: 1 Oct 2021 → 1 Oct 2021 |
Conference
Conference | 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 |
---|---|
By | Virtual, Online |
Periode | 01/10/2021 → 01/10/2021 |
Series | Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) |
---|---|
Volume | 13003 LNCS |
ISSN | 0302-9743 |
Bibliographical note
Publisher Copyright:
© 2021, Springer Nature Switzerland AG.
- 3D imaging, Deep learning, Segmentation, Sparse annotations, U-Net
Research areas
ID: 282672833