Semi-supervised Medical Image Segmentation via Learning Consistency Under Transformations
Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review
Standard
Semi-supervised Medical Image Segmentation via Learning Consistency Under Transformations. / Bortsova, Gerda; Dubost, Florian; Hogeweg, Laurens; Katramados, Ioannis; de Bruijne, Marleen.
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. 810-818 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), Vol. 11769 LNCS).Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review
Harvard
APA
Vancouver
Author
Bibtex
}
RIS
TY - GEN
T1 - Semi-supervised Medical Image Segmentation via Learning Consistency Under Transformations
AU - Bortsova, Gerda
AU - Dubost, Florian
AU - Hogeweg, Laurens
AU - Katramados, Ioannis
AU - de Bruijne, Marleen
PY - 2019/1/1
Y1 - 2019/1/1
N2 - The scarcity of labeled data often limits the application of supervised deep learning techniques for medical image segmentation. This has motivated the development of semi-supervised techniques that learn from a mixture of labeled and unlabeled images. In this paper, we propose a novel semi-supervised method that, in addition to supervised learning on labeled training images, learns to predict segmentations consistent under a given class of transformations on both labeled and unlabeled images. More specifically, in this work we explore learning equivariance to elastic deformations. We implement this through: (1) a Siamese architecture with two identical branches, each of which receives a differently transformed image, and (2) a composite loss function with a supervised segmentation loss term and an unsupervised term that encourages segmentation consistency between the predictions of the two branches. We evaluate the method on a public dataset of chest radiographs with segmentations of anatomical structures using 5-fold cross-validation. The proposed method reaches significantly higher segmentation accuracy compared to supervised learning. This is due to learning transformation consistency on both labeled and unlabeled images, with the latter contributing the most. We achieve the performance comparable to state-of-the-art chest X-ray segmentation methods while using substantially fewer labeled images.
AB - The scarcity of labeled data often limits the application of supervised deep learning techniques for medical image segmentation. This has motivated the development of semi-supervised techniques that learn from a mixture of labeled and unlabeled images. In this paper, we propose a novel semi-supervised method that, in addition to supervised learning on labeled training images, learns to predict segmentations consistent under a given class of transformations on both labeled and unlabeled images. More specifically, in this work we explore learning equivariance to elastic deformations. We implement this through: (1) a Siamese architecture with two identical branches, each of which receives a differently transformed image, and (2) a composite loss function with a supervised segmentation loss term and an unsupervised term that encourages segmentation consistency between the predictions of the two branches. We evaluate the method on a public dataset of chest radiographs with segmentations of anatomical structures using 5-fold cross-validation. The proposed method reaches significantly higher segmentation accuracy compared to supervised learning. This is due to learning transformation consistency on both labeled and unlabeled images, with the latter contributing the most. We achieve the performance comparable to state-of-the-art chest X-ray segmentation methods while using substantially fewer labeled images.
KW - Chest X-ray
KW - Segmentation
KW - Semi-supervised learning
U2 - 10.1007/978-3-030-32226-7_90
DO - 10.1007/978-3-030-32226-7_90
M3 - Article in proceedings
AN - SCOPUS:85075823024
SN - 9783030322250
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 810
EP - 818
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
T2 - 22nd International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2019
Y2 - 13 October 2019 through 17 October 2019
ER -
ID: 231952792