Superpixel-based and boundary-sensitive convolutional neural network for automated liver segmentation.
Superpixel-based and boundary-sensitive convolutional neural network for automated liver segmentation.
复制标题
DOI：
10.1088/1361-6560/aabd19
复制
发表时间：
2018-05-04
影响因子：
3.5
通讯作者：
Xing L
中科院分区：
文献类型：
--
作者：
Qin W;Wu J;Han F;Yuan Y;Zhao W;Ibragimov B;Gu J;Xing L
腹部CT中肝脏的分割是肝癌放射治疗计划的重要步骤。实际上，由于肝脏与其周围器官之间的低软组织对比度及其高度可变形的形状，肝脏的全自动分割仍然具有挑战性。这项工作的目的是开发一种新的基于超像素和边界敏感的卷积神经网络（SBBS-CNN）管道，用于自动肝脏分割。首先将整个CT图像划分为超像素区域，其中具有相似CT数的邻近像素被聚合。其次，我们通过将超像素标记为三类：内部肝脏、肝脏边界和非肝脏背景，将传统的二值分割转换为多项式分类。通过这样做，肝脏的边界区域被明确地识别并突出显示用于随后的分类。第三，我们为每个CT体积计算基于熵的显著性图，并利用该图来指导超像素上的图像块的采样。以这种方式，从信息区域（例如，具有不规则变化的肝脏边界），并且从均匀区域中提取较少的块。最后，构建并训练深度CNN管道来预测肝脏边界的概率图。我们在100名患者的队列中测试了所提出的算法。通过10倍交叉验证，SBBS-CNN实现了平均Dice相似系数为97.31±0.36%，平均对称表面距离为1.77± 0.49 mm。此外，它表现出上级的性能相比，最先进的方法，包括U-网络，基于像素的CNN，活动轮廓，水平集和图切割算法。SBBS-CNN为自动肝脏分割提供了准确有效的工具。还可以设想，所提出的框架是直接适用于其他医学图像分割的情况。
Segmentation of liver in abdominal computed tomography (CT) is an important step for radiation therapy planning of hepatocellular carcinoma. Practically, a fully automatic segmentation of liver remains challenging because of low soft tissue contrast between liver and its surrounding organs, and its highly deformable shape. The purpose of this work is to develop a novel superpixel-based and boundary sensitive convolutional neural network (SBBS-CNN) pipeline for automated liver segmentation. The entire CT images were first partitioned into superpixel regions, where nearby pixels with similar CT number were aggregated. Secondly, we converted the conventional binary segmentation into a multinomial classification by labeling the superpixels into three classes: interior liver, liver boundary, and non-liver background. By doing this, the boundary region of the liver was explicitly identified and highlighted for the subsequent classification. Thirdly, we computed an entropy-based saliency map for each CT volume, and leveraged this map to guide the sampling of image patches over the superpixels. In this way, more patches were extracted from informative regions (e.g., the liver boundary with irregular changes) and fewer patches were extracted from homogeneous regions. Finally, deep CNN pipeline was built and trained to predict the probability map of the liver boundary. We tested the proposed algorithm in a cohort of 100 patients. With 10-fold cross validation, the SBBS-CNN achieved mean Dice similarity coefficients of 97.31±0.36% and average symmetric surface distance of 1.77±0.49mm. Moreover, it showed superior performance in comparison with state-of-art methods, including U-Net, pixel-based CNN, active contour, level-sets and graph-cut algorithms. SBBS-CNN provides an accurate and effective tool for automated liver segmentation. It is also envisioned that the proposed framework is directly applicable in other medical image segmentation scenarios.
登录
查看更多内容
影响因子：
10.9
作者：
Hoogi A;Beaulieu CF;Cunha GM;Heba E;Sirlin CB;Napel S;Rubin DL
通讯作者：
Rubin DL
影响因子：
10.6
作者：
Guo Y;Gao Y;Shen D
通讯作者：
Shen D
影响因子：
64.8
作者：
Esteva A;Kuprel B;Novoa RA;Ko J;Swetter SM;Blau HM;Thrun S
通讯作者：
Thrun S
影响因子：
12
作者：
Mharib, Ahmed M.;Ramli, Abdul Rahman;Mahmood, Rozi Binti
通讯作者：
Mahmood, Rozi Binti
影响因子：
4.6
作者：
Ghafoorian M;Karssemeijer N;Heskes T;van Uden IWM;Sanchez CI;Litjens G;de Leeuw FE;van Ginneken B;Marchiori E;Platel B
通讯作者：
Platel B