A SEMI-SUPERVISED JOINT LEARNING APPROACH TO LEFT VENTRICULAR SEGMENTATION AND MOTION TRACKING IN ECHOCARDIOGRAPHY.
A SEMI-SUPERVISED JOINT LEARNING APPROACH TO LEFT VENTRICULAR SEGMENTATION AND MOTION TRACKING IN ECHOCARDIOGRAPHY.
复制标题
DOI：
10.1109/isbi45749.2020.9098664
复制
发表时间：
2020-04
期刊：
影响因子：
--
通讯作者：
Duncan JS
中科院分区：
文献类型：
--
作者：
Ta K;Ahn SS;Lu A;Stendahl JC;Sinusas AJ;Duncan JS
超声心动图的准确解释和分析对评估心血管健康非常重要。然而，运动跟踪往往依赖于心肌的准确分割，由于固有的超声特性，这很难获得。为了解决这一限制，我们提出了一种利用运动跟踪和分割中的重叠特征的半监督联合学习网络。该网络同时训练两个分支：一个用于运动跟踪，一个用于分割。每个分支都学习提取与各自任务相关的特征，并与其他分支共享。学习的运动估计通过时间传播手动分割的掩码，用于指导未来的分割预测。引入生理约束来加强真实的心脏行为。实验结果表明，人工合成和活体犬2D+t超声心动图序列在这两方面都优于一些竞争方法。
Accurate interpretation and analysis of echocardiography is important in assessing cardiovascular health. However, motion tracking often relies on accurate segmentation of the myocardium, which can be difficult to obtain due to inherent ultrasound properties. In order to address this limitation, we propose a semi-supervised joint learning network that exploits overlapping features in motion tracking and segmentation. The network simultaneously trains two branches: one for motion tracking and one for segmentation. Each branch learns to extract features relevant to their respective tasks and shares them with the other. Learned motion estimations propagate a manually segmented mask through time, which is used to guide future segmentation predictions. Physiological constraints are introduced to enforce realistic cardiac behavior. Experimental results on synthetic and in vivo canine 2D+t echocardiographic sequences outperform some competing methods in both tasks.