Bayesian uncertainty estimation for detection of long-tailed and unseen conditions in medical images.
Bayesian uncertainty estimation for detection of long-tailed and unseen conditions in medical images.
复制标题
用于检测医学图像中的长尾和不可见条件的贝叶斯不确定性估计。
DOI：
10.1117/1.jmi.10.5.054501
复制
发表时间：
2023
期刊：
影响因子：
--
通讯作者：
Yoshida,Hiroyuki
中科院分区：
文献类型：
--
作者：
Rezaei,Mina;Näppi,JanneJ;Bischl,Bernd;Yoshida,Hiroyuki
目的深度监督学习为开发各种计算机辅助诊断任务的稳健模型提供了一种有效的方法。然而，通常有一个基本假设，即训练数据集不同类别之间的样本频率要么相似，要么平衡。在现实世界的医学数据中，正类样本的出现频率往往太低，无法满足这一假设。因此，对能够自动识别和适应不平衡数据的现实世界条件的深度学习系统的需求尚未得到满足。方法我们提出了一种深度贝叶斯集成学习框架，以解决从医学图像训练时长尾和分布外（OOD）样本的表示学习问题。通过估计输入数据的相对不确定性，我们的框架可以适应不平衡数据以学习可概括的分类器。我们在四个公共医学成像数据集上训练和测试了我们的框架，这些数据集具有不同的不平衡比和成像模式，涵盖三个不同的学习任务：语义医学图像分割、OOD 检测和域内泛化。我们将我们的框架的性能与最先进的比较器方法的性能进行了比较。结果我们提出的框架在高分辨率 CT 和 MR 图像的语义分割以及 OOD 样本检测 (p< 0.01) 的所有性能指标 (pairwiset-test:p< 0.01) 中均显着优于比较器模型，从而在处理相关的长尾数据分布方面显示出显着的改进。域内泛化的结果还表明，我们的框架可以增强视网膜青光眼的预测，有助于临床决策过程。结论对所提出的深度贝叶斯集成学习框架进行动态蒙特卡洛丢失和损失组合的训练，可以对来自不同学习任务的不平衡医学成像数据集中的未见样本进行最佳泛化。
PurposeDeep supervised learning provides an effective approach for developing robust models for various computer-aided diagnosis tasks. However, there is often an underlying assumption that the frequencies of the samples between the different classes of the training dataset are either similar or balanced. In real-world medical data, the samples of positive classes often occur too infrequently to satisfy this assumption. Thus, there is an unmet need for deep-learning systems that can automatically identify and adapt to the real-world conditions of imbalanced data.ApproachWe propose a deep Bayesian ensemble learning framework to address the representation learning problem of long-tailed and out-of-distribution (OOD) samples when training from medical images. By estimating the relative uncertainties of the input data, our framework can adapt to imbalanced data for learning generalizable classifiers. We trained and tested our framework on four public medical imaging datasets with various imbalance ratios and imaging modalities across three different learning tasks: semantic medical image segmentation, OOD detection, and in-domain generalization. We compared the performance of our framework with those of state-of-the-art comparator methods.ResultsOur proposed framework outperformed the comparator models significantly across all performance metrics (pairwiset-test:p< 0.01) in the semantic segmentation of high-resolution CT and MR images as well as in the detection of OOD samples (p< 0.01), thereby showing significant improvement in handling the associated long-tailed data distribution. The results of the in-domain generalization also indicated that our framework can enhance the prediction of retinal glaucoma, contributing to clinical decision-making processes.ConclusionsTraining of the proposed deep Bayesian ensemble learning framework with dynamic Monte-Carlo dropout and a combination of losses yielded the best generalization to unseen samples from imbalanced medical imaging datasets across different learning tasks.