A Likelihood and Local Constraint Level Set Model for Liver Tumor Segmentation from CT Volumes
A Likelihood and Local Constraint Level Set Model for Liver Tumor Segmentation from CT Volumes
复制标题
DOI：
10.1109/tbme.2013.2267212
复制
发表时间：
2013-10-01
影响因子：
4.6
通讯作者：
Feng, David Dagan
中科院分区：
文献类型：
--
作者：
Li, Changyang;Wang, Xiuying;Feng, David Dagan
在肝脏肿瘤的计算机断层扫描中，常常存在密度不均匀，边界薄弱，并且肝脏肿瘤被具有相似密度的其他腹部结构包围。这些对准确的肝肿瘤分割造成了限制。我们提出了一个水平集模型结合似然能量与边缘能量。似然能量的最小化近似目标的密度分布和可以具有多个区域的背景的多峰密度分布。在边缘能量公式中，我们的边缘检测器保留了与弱边界的边缘相关的斜坡。我们将我们的方法与Chan-Vese和测地线水平集模型以及临床专家进行的手动分割进行了比较。Chan-Vese模型在肝肿瘤分割中并不成功，我们的模型优于测地线水平集模型。我们对18个临床数据集的结果表明，我们的算法的Jaccard距离误差为14.4 +/-5.3%，相对体积差为-8.1 +/-2.1%，平均表面距离为2.4 +/- 0.8 mm，RMS表面距离为2.9 +/- 0.7 mm，最大表面距离为7.2 +/- 3.1 mm。
In computed tomography of liver tumors there is often heterogeneous density, weak boundaries, and the liver tumors are surrounded by other abdominal structures with similar densities. These pose limitations to accurate the hepatic tumor segmentation. We propose a level set model incorporating likelihood energy with the edge energy. The minimization of the likelihood energy approximates the density distribution of the target and the multimodal density distribution of the background that can have multiple regions. In the edge energy formulation, our edge detector preserves the ramp associated with the edges for weak boundaries. We compared our approach to the Chan-Vese and the geodesic level set models and the manual segmentation performed by clinical experts. The Chan-Vese model was not successful in segmenting hepatic tumors and our model outperformed the geodesic level set model. Our results on 18 clinical datasets showed that our algorithm had a Jaccard distance error of 14.4 +/- 5.3%, relative volume difference of -8.1 +/- 2.1%, average surface distance of 2.4 +/- 0.8 mm, RMS surface distance of 2.9 +/- 0.7 mm, and the maximum surface distance of 7.2 +/- 3.1 mm.