In this section
@inproceedings{Shi:2018:10.1007/978-3-030-00937-3_65,
author = {Shi, Z and Zeng, G and Zhang, L and Zhuang, X and Li, L and Yang, G and Zheng, G},
doi = {10.1007/978-3-030-00937-3_65},
pages = {569--577},
title = {Bayesian VoxDRN: A Probabilistic Deep Voxelwise Dilated Residual Network for Whole Heart Segmentation from 3D MR Images},
url = {http://dx.doi.org/10.1007/978-3-030-00937-3_65},
year = {2018}
}
TY - CPAPER
AB - © 2018, Springer Nature Switzerland AG. In this paper, we propose a probabilistic deep voxelwise dilated residual network, referred as Bayesian VoxDRN, to segment the whole heart from 3D MR images. Bayesian VoxDRN can predict voxelwise class labels with a measure of model uncertainty, which is achieved by a dropout-based Monte Carlo sampling during testing to generate a posterior distribution of the voxel class labels. Our method has three compelling advantages. First, the dropout mechanism encourages the model to learn a distribution of weights with better data-explanation ability and prevents over-fitting. Second, focal loss and Dice loss are well encapsulated into a complementary learning objective to segment both hard and easy classes. Third, an iterative switch training strategy is introduced to alternatively optimize a binary segmentation task and a multi-class segmentation task for a further accuracy improvement. Experiments on the MICCAI 2017 multi-modality whole heart segmentation challenge data corroborate the effectiveness of the proposed method.
AU - Shi,Z
AU - Zeng,G
AU - Zhang,L
AU - Zhuang,X
AU - Li,L
AU - Yang,G
AU - Zheng,G
DO - 10.1007/978-3-030-00937-3_65
EP - 577
PY - 2018///
SN - 0302-9743
SP - 569
TI - Bayesian VoxDRN: A Probabilistic Deep Voxelwise Dilated Residual Network for Whole Heart Segmentation from 3D MR Images
UR - http://dx.doi.org/10.1007/978-3-030-00937-3_65
ER -
For enquiries about the MRI Physics Initiative, please contact:
Senior MR Physicist
Mary Finnegan
Imperial Research Fellow
Matthew Grech-Sollars
BRC MR Physics Fellow
Pete Lally