Adams Jadie, Elhabian Shireen Y
Scientific Computing and Imaging Institute, University of Utah, UT, USA.
Kahlert School of Computing, University of Utah, UT, USA.
Med Image Comput Comput Assist Interv. 2023 Oct;14222:346-356. doi: 10.1007/978-3-031-43898-1_34. Epub 2023 Oct 1.
Statistical shape modeling (SSM) enables population-based quantitative analysis of anatomical shapes, informing clinical diagnosis. Deep learning approaches predict correspondence-based SSM directly from unsegmented 3D images but require calibrated uncertainty quantification, motivating Bayesian formulations. Variational information bottleneck DeepSSM (VIB-DeepSSM) is an effective, principled framework for predicting probabilistic shapes of anatomy from images with aleatoric uncertainty quantification. However, VIB is only half-Bayesian and lacks epistemic uncertainty inference. We derive a fully Bayesian VIB formulation and demonstrate the efficacy of two scalable implementation approaches: concrete dropout and batch ensemble. Additionally, we introduce a novel combination of the two that further enhances uncertainty calibration via multimodal marginalization. Experiments on synthetic shapes and left atrium data demonstrate that the fully Bayesian VIB network predicts SSM from images with improved uncertainty reasoning without sacrificing accuracy.
统计形状建模（SSM）能够对解剖形状进行基于人群的定量分析，为临床诊断提供依据。深度学习方法可直接从未分割的3D图像预测基于对应关系的SSM，但需要校准不确定性量化，这推动了贝叶斯公式的发展。变分信息瓶颈深度SSM（VIB-DeepSSM）是一个有效的、有原则的框架，用于从具有随机不确定性量化的图像中预测解剖结构的概率形状。然而，VIB只是半贝叶斯的，缺乏认知不确定性推理。我们推导了一个完全贝叶斯的VIB公式，并展示了两种可扩展实现方法的有效性：具体失活和批集成。此外，我们引入了两者的新颖组合，通过多模态边缘化进一步增强不确定性校准。对合成形状和左心房数据的实验表明，完全贝叶斯VIB网络能够从图像中预测SSM，在不牺牲准确性的情况下改善不确定性推理。
