TY - JOUR
T1 - BIRNet
T2 - Brain image registration using dual-supervised fully convolutional networks
AU - Fan, Jingfan
AU - Cao, Xiaohuan
AU - Yap, Pew Thian
AU - Shen, Dinggang
N1 - Funding Information:
This work was supported in part by NIH grants (EB006733, EB008374, MH100217, MH108914, AG041721, AG049371, AG042599, AG053867, EB022880, MH110274).
Funding Information:
This work was supported in part by NIH grants ( EB006733 , EB008374 , MH100217 , MH108914 , AG041721 , AG049371 , AG042599 , AG053867 , EB022880 , MH110274 ).
Publisher Copyright:
© 2019 Elsevier B.V.
PY - 2019/5
Y1 - 2019/5
N2 - In this paper, we propose a deep learning approach for image registration by predicting deformation from image appearance. Since obtaining ground-truth deformation fields for training can be challenging, we design a fully convolutional network that is subject to dual-guidance: (1) Ground-truth guidance using deformation fields obtained by an existing registration method; and (2) Image dissimilarity guidance using the difference between the images after registration. The latter guidance helps avoid overly relying on the supervision from the training deformation fields, which could be inaccurate. For effective training, we further improve the deep convolutional network with gap filling, hierarchical loss, and multi-source strategies. Experiments on a variety of datasets show promising registration accuracy and efficiency compared with state-of-the-art methods.
AB - In this paper, we propose a deep learning approach for image registration by predicting deformation from image appearance. Since obtaining ground-truth deformation fields for training can be challenging, we design a fully convolutional network that is subject to dual-guidance: (1) Ground-truth guidance using deformation fields obtained by an existing registration method; and (2) Image dissimilarity guidance using the difference between the images after registration. The latter guidance helps avoid overly relying on the supervision from the training deformation fields, which could be inaccurate. For effective training, we further improve the deep convolutional network with gap filling, hierarchical loss, and multi-source strategies. Experiments on a variety of datasets show promising registration accuracy and efficiency compared with state-of-the-art methods.
KW - Brain MR image
KW - Convolutional neural networks
KW - Hierarchical registration
KW - Image registration
UR - http://www.scopus.com/inward/record.url?scp=85063570767&partnerID=8YFLogxK
U2 - 10.1016/j.media.2019.03.006
DO - 10.1016/j.media.2019.03.006
M3 - Article
C2 - 30939419
AN - SCOPUS:85063570767
SN - 1361-8415
VL - 54
SP - 193
EP - 206
JO - Medical Image Analysis
JF - Medical Image Analysis
ER -