LungRegNet: An unsupervised deformable image registration method for 4D-CT lung
LungRegNet: An unsupervised deformable image registration method for 4D-CT lung
复制标题
DOI：
10.1002/mp.14065
复制
发表时间：
2020-02-26
期刊：
影响因子：
3.8
通讯作者：
Yang, Xiaofeng
中科院分区：
文献类型：
--
作者：
Fu, Yabo;Lei, Yang;Yang, Xiaofeng
目的研究一种准确、快速的四维CT肺部图像形变配准方法。基于深度学习的方法有可能在一些前向预测中快速预测变形矢量场（DVF）。我们开发了一种用于4D-CT肺部DIR的无监督深度学习方法，在配准准确性、鲁棒性和计算速度方面具有出色的性能。方法利用深度学习提出了一种快速准确的4D-CT肺部DIR方法，即LungRegNet。LungRegNet由两个子网组成，分别是CoarseNet和FineNet。顾名思义，CoarseNet在粗尺度图像上预测大的肺部运动，而FineNet在细尺度图像上预测局部肺部运动。CoarseNet和FineNet都包括一个生成器和一个编译器。生成器被训练为直接预测DVF以使运动图像变形。训练神经网络以区分变形图像和原始图像。CoarseNet首先被训练来变形移动图像。然后FineNet使用变形的图像进行FineNet训练。为了提高配准精度的LungRegNet，我们生成血管增强图像生成肺血管概率图之前的网络prediction.Results我们进行了五重交叉验证10个4D-CT数据集从我们的部门。为了与其他方法进行比较，我们还使用单独的10个DIRLAB数据集测试了我们的方法，每个数据集提供300个手动地标对，用于目标配准误差（TRE）计算。我们的研究结果表明，LungRegNet在TRE方面比DIRLAB数据集文献中其他基于深度学习的方法实现了更好的配准精度。与传统的肺血管增强方法相比，LungRegNet可以产生相当的配准精度，TRE小于2 mm。将肺血管增强和肺血管增强集成到网络中对于获得4D-CT肺血管增强的高配准精度至关重要。在我们的数据集和DIRLAB数据集上，TRE的平均值和标准差分别为1.00 +/- 0.53 mm和1.59 +/- 1.58 mm.Conclusions已经开发了一种基于无监督深度学习的方法来快速准确地配准4D-CT肺部图像。LungRegNet的表现优于基于深度学习的同行，并在TRE方面实现了出色的配准精度。
Purpose To develop an accurate and fast deformable image registration (DIR) method for four-dimensional computed tomography (4D-CT) lung images. Deep learning-based methods have the potential to quickly predict the deformation vector field (DVF) in a few forward predictions. We have developed an unsupervised deep learning method for 4D-CT lung DIR with excellent performances in terms of registration accuracies, robustness, and computational speed.Methods A fast and accurate 4D-CT lung DIR method, namely LungRegNet, was proposed using deep learning. LungRegNet consists of two subnetworks which are CoarseNet and FineNet. As the name suggests, CoarseNet predicts large lung motion on a coarse scale image while FineNet predicts local lung motion on a fine scale image. Both the CoarseNet and FineNet include a generator and a discriminator. The generator was trained to directly predict the DVF to deform the moving image. The discriminator was trained to distinguish the deformed images from the original images. CoarseNet was first trained to deform the moving images. The deformed images were then used by the FineNet for FineNet training. To increase the registration accuracy of the LungRegNet, we generated vessel-enhanced images by generating pulmonary vasculature probability maps prior to the network prediction.Results We performed fivefold cross validation on ten 4D-CT datasets from our department. To compare with other methods, we also tested our method using separate 10 DIRLAB datasets that provide 300 manual landmark pairs per case for target registration error (TRE) calculation. Our results suggested that LungRegNet has achieved better registration accuracy in terms of TRE than other deep learning-based methods available in the literature on DIRLAB datasets. Compared to conventional DIR methods, LungRegNet could generate comparable registration accuracy with TRE smaller than 2 mm. The integration of both the discriminator and pulmonary vessel enhancements into the network was crucial to obtain high registration accuracy for 4D-CT lung DIR. The mean and standard deviation of TRE were 1.00 +/- 0.53 mm and 1.59 +/- 1.58 mm on our datasets and DIRLAB datasets respectively.Conclusions An unsupervised deep learning-based method has been developed to rapidly and accurately register 4D-CT lung images. LungRegNet has outperformed its deep-learning-based peers and achieved excellent registration accuracy in terms of TRE.