Accurate Segmentation of Heart Volume in CTA With Landmark-Based Registration and Fully Convolutional Network
Accurate Segmentation of Heart Volume in CTA With Landmark-Based Registration and Fully Convolutional Network
复制标题
利用基于地标的配准和全卷积网络在 CTA 中准确分割心脏体积
DOI：
10.1109/access.2019.2912467
复制
发表时间：
2019-04
期刊：
影响因子：
3.9
通讯作者：
Hou Yuqing
中科院分区：
文献类型：
--
作者：
Zhao Fengjun;Hu Haowen;Chen Yibing;Liang Jimin;He Xiaowei;Hou Yuqing
CTA图像中心脏结构的准确描绘仍然具有挑战性，因为不同成分的外观相似，个体之间的形状和强度差异很大，特别是心脏体积和背景(包括脂肪组织、大血管和肝脏)的附近。因此，我们提出了一种基于标志点配准(LMR)和3D完全卷积网络(3D-FCN)的心脏体积精确分割方法。首先在训练图像中定义均匀约束的地标，然后训练回归森林模型(RFM)在测试图像中检测这些地标。其次，利用浅层神经网络对训练图像和测试图像中的地标进行配准，引导标签从地图集传播到测试图像。经过多数投票的标记融合后，我们最终构造了3D-FCN来进一步细化低投票值的边界体素。在22幅心脏CTA图像中，我们将该方法与多图谱分割、主动形状模型、DeepMedic、LMR+DeepMedic以及分别实现的LMR和3D-FCN进行了比较。实验结果表明，该方法分割心脏体积的平均准确率为96.25%，Dice系数为93.98%，平均Hausdorff距离为2.12个体素。此外，通过训练时间和收敛损失的比较，构建的3D-FCN的训练效率高于DeepMEDIC。所提出的心脏分割不仅可以提供四个心腔的准确感兴趣区(ROI)，而且可以提供主要血管和冠状动脉的感兴趣区。
Accurate delineation of cardiac structures in CTA images remains challenging, due to the similar appearance of different components, the highly variable shape and intensity among individuals, and especially the vicinity of heart volumes and backgrounds (including fat tissues, great vessels, and livers). Therefore, we proposed an accurate heart volume segmentation method with landmark-based registration (LMR) and 3D fully convolutional network (3D-FCN). First, we defined uniformity constrained landmarks in the training images and then trained a regression forest model (RFM) to detect these landmarks in the testing image. Second, the registration between the landmarks in each training image and the test image was performed by a shallow neural network, which guided the label propagation from atlases to the test image. After the label fusion with majority voting, we finally constructed a 3D-FCN to further refine the boundary voxels with low voting values. In 22 cardiac CTA images, we compared our method with multi-atlas segmentation, active shape model, DeepMedic, LMR + DeepMedic, and the separately implemented LMR and 3D-FCN. The results demonstrated the superiority of our method for the segmentation of heart volumes, with the average accuracy, dice coefficient, and mean Hausdorff distance as 96.25%, 93.98%, and 2.12 voxels, respectively. Furthermore, the training efficiency of the constructed 3D-FCN was higher than that of the DeepMedic through the comparison of training time and convergence of loss. The proposed heart segmentation could provide an accurate region of interest (ROI) for not only the four heart chambers but also the major trunks of vessels and coronary arteries.
登录
查看更多内容
使用基于非线性场的模型进行心脏分割的动态更新图集
DOI：
10.1002/rcs.1785
发表时间：
2016
影响因子：
2.5
作者：
Cai Ken;Yang Rongqian;Yue Hongwei;Li Lihua;Ou Shanxing;Liu Feng
通讯作者：
Liu Feng
影响因子：
19.7
作者：
Rubin GD;Leipsic J;Joseph Schoepf U;Fleischmann D;Napel S
通讯作者：
Napel S
DOI：
10.1117/12.2293811
发表时间：
2018-03
期刊：
--
影响因子：
--
作者：
Hui Tang;Mehdi Moradi;A. Harouni;Hongzhi Wang;Gopalkrishna Veni;Prasanth Prasanna;T. Syeda-Mahmood
通讯作者：
Hui Tang;Mehdi Moradi;A. Harouni;Hongzhi Wang;Gopalkrishna Veni;Prasanth Prasanna;T. Syeda-Mahmood
影响因子：
--
作者：
Fei Ma;Jiquan Liu;Bin Wang-;H. Duan
通讯作者：
Fei Ma;Jiquan Liu;Bin Wang-;H. Duan
影响因子：
10.6
作者：
Zhang J;Gao Y;Gao Y;Munsell BC;Shen D
通讯作者：
Shen D