TY - GEN
T1 - Multi-task learning for neonatal brain segmentation using 3d dense-unet with dense attention guided by geodesic distance
AU - Bui, Toan Duc
AU - Wang, Li
AU - Chen, Jian
AU - Lin, Weili
AU - Li, Gang
AU - Shen, Dinggang
N1 - Funding Information:
Acknowledgment. This work was supported in part by NIH Grants MH107815, MH109773, MH116225, and MH117943.
Publisher Copyright:
© Springer Nature Switzerland AG 2019.
PY - 2019
Y1 - 2019
N2 - The deep convolutional neural network has achieved outstanding performance on neonatal brain MRI tissue segmentation. However, it may fail to produce reasonable results on unseen datasets that have different imaging appearance distributions with the training data. The main reason is that deep learning models tend to have a good fitting to the training dataset, but do not lead to a good generalization on the unseen datasets. To address this problem, we propose a multi-task learning method, which simultaneously learns both tissue segmentation and geodesic distance regression to regularize a shared encoder network. Furthermore, a dense attention gate is explored to force the network to learn rich contextual information. By using three neonatal brain datasets with different imaging protocols from different scanners, our experimental results demonstrate superior performance of our proposed method over the existing deep learning-based methods on the unseen datasets.
AB - The deep convolutional neural network has achieved outstanding performance on neonatal brain MRI tissue segmentation. However, it may fail to produce reasonable results on unseen datasets that have different imaging appearance distributions with the training data. The main reason is that deep learning models tend to have a good fitting to the training dataset, but do not lead to a good generalization on the unseen datasets. To address this problem, we propose a multi-task learning method, which simultaneously learns both tissue segmentation and geodesic distance regression to regularize a shared encoder network. Furthermore, a dense attention gate is explored to force the network to learn rich contextual information. By using three neonatal brain datasets with different imaging protocols from different scanners, our experimental results demonstrate superior performance of our proposed method over the existing deep learning-based methods on the unseen datasets.
KW - Attention
KW - Geodesic distance
KW - Multi-task learning
KW - Neonatal brain segmentation
UR - http://www.scopus.com/inward/record.url?scp=85075659372&partnerID=8YFLogxK
U2 - 10.1007/978-3-030-33391-1_28
DO - 10.1007/978-3-030-33391-1_28
M3 - Conference contribution
AN - SCOPUS:85075659372
SN - 9783030333904
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 243
EP - 251
BT - Domain Adaptation and Representation Transfer and Medical Image Learning with Less Labels and Imperfect Data First MICCAI Workshop, DART 2019 and First International Workshop, MIL3ID 2019 Shenzhen, Held in Conjunction with MICCAI 2019 Shenzhen, 2019 Proceedings
A2 - Wang, Qian
A2 - Milletari, Fausto
A2 - Rieke, Nicola
A2 - Nguyen, Hien V.
A2 - Roysam, Badri
A2 - Albarqouni, Shadi
A2 - Cardoso, M. Jorge
A2 - Xu, Ziyue
A2 - Kamnitsas, Konstantinos
A2 - Patel, Vishal
A2 - Jiang, Steve
A2 - Zhou, Kevin
A2 - Luu, Khoa
A2 - Le, Ngan
PB - Springer
T2 - 1st MICCAI Workshop on Domain Adaptation and Representation Transfer, DART 2019, and the 1st International Workshop on Medical Image Learning with Less Labels and Imperfect Data, MIL3ID 2019, held in conjunction with 22nd International Conference on Medical Image Computing and Computer- Assisted Intervention, MICCAI 2019
Y2 - 13 October 2019 through 17 October 2019
ER -