Adversarial learning for mono- or multi-modal registration
Adversarial learning for mono- or multi-modal registration
复制标题
DOI：
10.1016/j.media.2019.101545
复制
发表时间：
2019-12-01
影响因子：
10.9
通讯作者：
Shen, Dinggang
中科院分区：
文献类型：
--
作者：
Fan, Jingfan;Cao, Xiaohuan;Shen, Dinggang
本文介绍了一种用于图像配准的无监督对抗相似性网络。与现有的深度学习配准方法不同，我们的方法可以训练可变形的配准网络，而不需要地面真实变形和特定的相似性度量。我们用一个可变形的变换层连接一个配准网络和一个判别网络。配准网络的训练与判别网络的反馈，这是设计来判断一对注册的图像是否足够相似。使用对抗训练，配准网络被训练成预测足够准确以欺骗辨别网络的变形。因此，所提出的方法是一个通用的配准框架，它可以适用于单模态和多模态图像配准。在四个脑MRI数据集和多模态骨盆图像数据集上的实验表明，与最先进的配准方法（包括基于深度学习的配准方法）相比，我们的方法在准确性，效率和通用性方面具有良好的配准性能。(C)2019 Elsevier B.V.版权所有。
This paper introduces an unsupervised adversarial similarity network for image registration. Unlike existing deep learning registration methods, our approach can train a deformable registration network without the need of ground-truth deformations and specific similarity metrics. We connect a registration network and a discrimination network with a deformable transformation layer. The registration network is trained with the feedback from the discrimination network, which is designed to judge whether a pair of registered images are sufficiently similar. Using adversarial training, the registration network is trained to predict deformations that are accurate enough to fool the discrimination network. The proposed method is thus a general registration framework, which can be applied for both mono-modal and multi-modal image registration. Experiments on four brain MRI datasets and a multi-modal pelvic image dataset indicate that our method yields promising registration performance in accuracy, efficiency and generalizability compared with state-of-the-art registration methods, including those based on deep learning. (C) 2019 Elsevier B.V. All rights reserved.