Unsupervised Reverse Domain Adaptation for Synthetic Medical Images via Adversarial Training
Unsupervised Reverse Domain Adaptation for Synthetic Medical Images via Adversarial Training
复制标题
DOI：
10.1109/tmi.2018.2842767
复制
发表时间：
2018-12-01
影响因子：
10.6
通讯作者：
Durr, Nicholas J.
中科院分区：
文献类型：
--
作者：
Mahmood, Faisal;Chen, Richard;Durr, Nicholas J.
为了实现深度学习在医学成像中的全部潜力，需要大型注释数据集进行训练。由于隐私问题，缺乏可用于注释的专家，罕见疾病的代表性不足以及标准化差，这些数据集很难获得。在传统的视觉应用中，已经解决了注释数据的缺乏问题，这些应用使用经由无监督对抗训练细化的合成图像来看起来像真实的图像。然而，由于在真实的人体组织中发现的复杂多样的特征集，这种方法难以扩展到一般的医学成像。我们提出了一个新的框架，使用反向流，其中对抗训练用于使真实的医学图像更像合成图像，并通过自正则化保留临床相关的功能。然后，这些领域适应的合成图像可以通过在合成医学图像的大型数据集上训练的网络进行准确解释。我们实现这种方法上臭名昭著的困难任务的深度估计从单目内窥镜，它有各种应用在结肠镜检查，机器人手术，和侵入性内窥镜手术。我们在使用内窥镜和解剖学上真实的结肠的精确前向模型生成的合成图像的大数据集上训练深度估计器。我们的分析表明，内窥镜深度估计的结构相似性在一个真实的猪结肠预测从一个网络训练的合成数据单独使用反向域自适应提高了78.7%。
To realize the full potential of deep learning for medical imaging, large annotated datasets are required for training. Such datasets are difficult to acquire due to privacy issues, lack of experts available for annotation, underrepresentation of rare conditions, and poor standardization. The lack of annotated data has been addressed in conventional vision applications using synthetic images refined via unsupervised adversarial training to look like real images. However, this approach is difficult to extend to general medical imaging because of the complex and diverse set of features found in real human tissues. We propose a novel framework that uses a reverse flow, where adversarial training is used to make real medical images more like synthetic images, and clinically-relevant features are preserved via self-regularization. These domain-adapted synthetic-like images can then be accurately interpreted by networks trained on large datasets of synthetic medical images. We implement this approach on the notoriously difficult task of depth-estimation from monocular endoscopy which has a variety of applications in colonoscopy, robotic surgery, and invasive endoscopic procedures. We train a depth estimator on a large data set of synthetic images generated using an accurate forward model of an endoscope and an anatomically-realistic colon. Our analysis demonstrates that the structural similarity of endoscopy depth estimation in a real pig colon predicted from a network trained solely on synthetic data improved by 78.7% by using reverse domain adaptation.