Chair of Cellular and Molecular Imaging, Comprehensive Heart Failure Center (CHFC), University Hospital Würzburg, Am Schwarzenberg 15, 97078, Würzburg, Germany.
BMC Med Imaging. 2021 Feb 15;21(1):27. doi: 10.1186/s12880-021-00551-1.
Image segmentation is a common task in medical imaging e.g., for volumetry analysis in cardiac MRI. Artificial neural networks are used to automate this task with performance similar to manual operators. However, this performance is only achieved in the narrow tasks networks are trained on. Performance drops dramatically when data characteristics differ from the training set properties. Moreover, neural networks are commonly considered black boxes, because it is hard to understand how they make decisions and why they fail. Therefore, it is also hard to predict whether they will generalize and work well with new data. Here we present a generic method for segmentation model interpretation. Sensitivity analysis is an approach where model input is modified in a controlled manner and the effect of these modifications on the model output is evaluated. This method yields insights into the sensitivity of the model to these alterations and therefore to the importance of certain features on segmentation performance.
We present an open-source Python library (misas), that facilitates the use of sensitivity analysis with arbitrary data and models. We show that this method is a suitable approach to answer practical questions regarding use and functionality of segmentation models. We demonstrate this in two case studies on cardiac magnetic resonance imaging. The first case study explores the suitability of a published network for use on a public dataset the network has not been trained on. The second case study demonstrates how sensitivity analysis can be used to evaluate the robustness of a newly trained model.
Sensitivity analysis is a useful tool for deep learning developers as well as users such as clinicians. It extends their toolbox, enabling and improving interpretability of segmentation models. Enhancing our understanding of neural networks through sensitivity analysis also assists in decision making. Although demonstrated only on cardiac magnetic resonance images this approach and software are much more broadly applicable.
图像分割是医学成像中的一项常见任务，例如在心脏 MRI 的容积分析中。人工神经网络可用于自动执行此任务，其性能与手动操作相当。然而，只有在网络接受过训练的狭窄任务中才能实现这种性能。当数据特征与训练集属性不同时，性能会急剧下降。此外，神经网络通常被认为是黑盒，因为很难理解它们是如何做出决策的，以及为什么它们会失败。因此，也很难预测它们是否会推广并在新数据上正常工作。在这里，我们提出了一种用于分割模型解释的通用方法。敏感性分析是一种方法，其中模型输入以受控的方式进行修改，并评估这些修改对模型输出的影响。这种方法可以深入了解模型对这些改变的敏感性，从而了解某些特征对分割性能的重要性。
我们提出了一个开源的 Python 库（misas），它可以方便地使用敏感性分析来处理任意数据和模型。我们表明，这种方法是回答有关分割模型使用和功能的实际问题的合适方法。我们在两个心脏磁共振成像的案例研究中证明了这一点。第一个案例研究探讨了将已发布的网络用于该网络未接受过训练的公共数据集的适用性。第二个案例研究演示了如何使用敏感性分析来评估新训练模型的鲁棒性。
敏感性分析是深度学习开发人员以及临床医生等用户的有用工具。它扩展了他们的工具包，使分割模型的可解释性得到增强和改善。通过敏感性分析增强我们对神经网络的理解还有助于决策制定。尽管仅在心脏磁共振图像上进行了演示，但这种方法和软件具有更广泛的适用性。
