Evaluating reinforcement learning agents for anatomical landmark detection.

Evaluating reinforcement learning agents for anatomical landmark detection.
复制标题

DOI:
10.1016/j.media.2019.02.007
复制
发表时间:
2019-04
影响因子:
10.9
通讯作者:
Rueckert D
Rueckert D
中科院分区:
工程技术1区
文献类型:
--
作者:
Alansary A;Oktay O;Li Y;Folgoc LL;Hou B;Vaillant G;Kamnitsas K;Vlontzos A;Glocker B;Kainz B;Rueckert D

文献摘要

参考文献

被引文献

相似文献

解剖标志的自动检测是医学图像分析中广泛应用的一个重要步骤。手动标注地标是一项繁琐的工作,容易产生观察者错误。在本文中,我们评估了新的深度强化学习(RL)策略来训练智能体,这些智能体可以精确和鲁棒地定位医学扫描中的目标地标。人工强化学习代理通过与环境(在我们的例子中是3D图像)交互来学习识别到地标的最佳路径。此外,我们研究了固定尺度和多尺度搜索策略的使用,并以一种从粗到细的方式采用了新的分层动作步骤。使用三种不同的医学成像数据集:胎儿头部超声(US)、成人大脑和心脏磁共振成像(MRI),评估了几种深度q -网络(DQN)架构用于检测多个地标。我们的代理的性能超过了最先进的监督和强化学习方法。我们的实验还表明,在具有大视场和噪声背景的图像(如心脏MRI)中,多尺度搜索策略明显优于固定尺度代理。此外,新的分层步骤可以显著加快搜索过程,速度是原来的4 - 5倍。
Automatic detection of anatomical landmarks is an important step for a wide range of applications in medical image analysis. Manual annotation of landmarks is a tedious task and prone to observer errors. In this paper, we evaluate novel deep reinforcement learning (RL) strategies to train agents that can precisely and robustly localize target landmarks in medical scans. An artificial RL agent learns to identify the optimal path to the landmark by interacting with an environment, in our case 3D images. Furthermore, we investigate the use of fixed- and multiscale search strategies with novel hierarchical action steps in a coarse-to-fine manner. Several deep Q-network (DQN) architectures are evaluated for detecting multiple landmarks using three different medical imaging datasets: fetal head ultrasound (US), adult brain and cardiac magnetic resonance imaging (MRI). The performance of our agents surpasses state-of-the-art supervised and RL methods. Our experiments also show that multi-scale search strategies perform significantly better than fixed-scale agents in images with large field of view and noisy background such as in cardiac MRI. Moreover, the novel hierarchical steps can significantly speed up the searching process by a factor of 4 − 5 times.
DOI: 10.1186/1532-429x-16-16
发表时间: 2014-02-03
影响因子: 6.4
作者:
de Marvao, Antonio;Dawes, Timothy J. W.;O'Regan, Declan P.
通讯作者: O'Regan, Declan P.
DOI: 10.1016/j.media.2013.01.001
发表时间: 2013-12-01
影响因子: 10.9
作者:
Criminisi, A.;Robertson, D.;Siddiqui, K.
通讯作者: Siddiqui, K.
DOI: 10.1016/j.neuroimage.2009.02.030
发表时间: 2009-07-01
期刊: NEUROIMAGE
影响因子: 5.7
作者:
Ardekani, Babak A.;Bachman, Alvin H.
通讯作者: Bachman, Alvin H.
DOI: 10.1038/nature14236
发表时间: 2015-02-26
期刊: NATURE
影响因子: 64.8
作者:
Mnih, Volodymyr;Kavukcuoglu, Koray;Hassabis, Demis
通讯作者: Hassabis, Demis
DOI: 10.1016/j.media.2015.04.007
发表时间: 2015-07-01
影响因子: 10.9
作者:
Gauriau, Romane;Cuingnet, Remi;Bloch, Isabelle
通讯作者: Bloch, Isabelle