One Network to Segment Them All: A General, Lightweight System for Accurate 3D Medical Image Segmentation
Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review
Standard
One Network to Segment Them All : A General, Lightweight System for Accurate 3D Medical Image Segmentation. / Perslev, Mathias; Dam, Erik Bjørnager; Pai, Akshay; Igel, Christian.
Medical Image Computing and Computer Assisted Intervention – MICCAI 2019 - 22nd International Conference, Proceedings. ed. / Dinggang Shen; Pew-Thian Yap; Tianming Liu; Terry M. Peters; Ali Khan; Lawrence H. Staib; Caroline Essert; Sean Zhou. Springer VS, 2019. p. 30-38 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), Vol. 11765 LNCS).Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review
Harvard
APA
Vancouver
Author
Bibtex
}
RIS
TY - GEN
T1 - One Network to Segment Them All
T2 - 22nd International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2019
AU - Perslev, Mathias
AU - Dam, Erik Bjørnager
AU - Pai, Akshay
AU - Igel, Christian
PY - 2019
Y1 - 2019
N2 - Many recent medical segmentation systems rely on powerful deep learning models to solve highly specific tasks. To maximize performance, it is standard practice to evaluate numerous pipelines with varying model topologies, optimization parameters, pre- & postprocessing steps, and even model cascades. It is often not clear how the resulting pipeline transfers to different tasks. We propose a simple and thoroughly evaluated deep learning framework for segmentation of arbitrary medical image volumes. The system requires no task-specific information, no human interaction and is based on a fixed model topology and a fixed hyperparameter set, eliminating the process of model selection and its inherent tendency to cause method-level over-fitting. The system is available in open source and does not require deep learning expertise to use. Without task-specific modifications, the system performed better than or similar to highly specialized deep learning methods across 3 separate segmentation tasks. In addition, it ranked 5-th and 6-th in the first and second round of the 2018 Medical Segmentation Decathlon comprising another 10 tasks. The system relies on multi-planar data augmentation which facilitates the application of a single 2D architecture based on the familiar U-Net. Multi-planar training combines the parameter efficiency of a 2D fully convolutional neural network with a systematic train- and test-time augmentation scheme, which allows the 2D model to learn a representation of the 3D image volume that fosters generalization.
AB - Many recent medical segmentation systems rely on powerful deep learning models to solve highly specific tasks. To maximize performance, it is standard practice to evaluate numerous pipelines with varying model topologies, optimization parameters, pre- & postprocessing steps, and even model cascades. It is often not clear how the resulting pipeline transfers to different tasks. We propose a simple and thoroughly evaluated deep learning framework for segmentation of arbitrary medical image volumes. The system requires no task-specific information, no human interaction and is based on a fixed model topology and a fixed hyperparameter set, eliminating the process of model selection and its inherent tendency to cause method-level over-fitting. The system is available in open source and does not require deep learning expertise to use. Without task-specific modifications, the system performed better than or similar to highly specialized deep learning methods across 3 separate segmentation tasks. In addition, it ranked 5-th and 6-th in the first and second round of the 2018 Medical Segmentation Decathlon comprising another 10 tasks. The system relies on multi-planar data augmentation which facilitates the application of a single 2D architecture based on the familiar U-Net. Multi-planar training combines the parameter efficiency of a 2D fully convolutional neural network with a systematic train- and test-time augmentation scheme, which allows the 2D model to learn a representation of the 3D image volume that fosters generalization.
U2 - 10.1007/978-3-030-32245-8_4
DO - 10.1007/978-3-030-32245-8_4
M3 - Article in proceedings
AN - SCOPUS:85075675727
SN - 9783030322441
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 30
EP - 38
BT - Medical Image Computing and Computer Assisted Intervention – MICCAI 2019 - 22nd International Conference, Proceedings
A2 - Shen, Dinggang
A2 - Yap, Pew-Thian
A2 - Liu, Tianming
A2 - Peters, Terry M.
A2 - Khan, Ali
A2 - Staib, Lawrence H.
A2 - Essert, Caroline
A2 - Zhou, Sean
PB - Springer VS
Y2 - 13 October 2019 through 17 October 2019
ER -
ID: 237709157