Multimodal vs. unimodal approaches to uncertainty in 3D image segmentation under distribution shifts

Machine learning has been widely adopted across sectors, yet its application in medical imaging remains challenging due to distribution shifts in real-world data. Deployed models often encounter samples that differ from the training dataset, particularly in the health domain, leading to performance issues. This limitation hinders the expressiveness and reliability of deep learning models in health applications. Thus, it becomes crucial to identify methods capable of producing reliable uncertainty estimation in the context of distribution shifts in the health sector. In this paper, we explore the feasibility of using cutting-edge Bayesian and non-Bayesian methods to detect distributionally shifted samples, aiming to achieve reliable and trustworthy diagnostic predictions in segmentation task. Specifically, we compare three distinct uncertainty estimation methods, each designed to capture either unimodal or multimodal aspects in the posterior distribution. Our findings demonstrate that methods capable of addressing multimodal characteristics in the posterior distribution, offer more dependable uncertainty estimates. This research contributes to enhancing the utility of deep learning in healthcare, making diagnostic predictions more robust and trustworthy. • Distributional shifts can significantly drop performance in the medical domain. • Reliable uncertainty estimation is vital for safe deployment of deep learning models in healthcare, enabling more rational and informed decesion-making. • Not all uncertainty estimation methods produce reliable results under distribution shifts. • Characterizing methods that produce reliable uncertainties is crucial in medical imaging. • Methods capturing multiple posterior modes offer more accurate uncertainty estimates under distribution shifts.