Uncertainty-Error correlations in Evidential Deep Learning models for biomedical segmentation
This work addresses uncertainty quantification for biomedical segmentation, which is incremental as it builds on existing EDL frameworks to improve error detection in critical medical tasks.
The study applied Evidential Deep Learning (EDL) to biomedical image segmentation, finding that EDL models with U-Net backbones achieved higher correlations between prediction errors and uncertainties compared to baseline methods like Shannon entropy, Monte-Carlo Dropout, and Deep Ensembles, and also performed better in active learning with similar Dice-Sorensen coefficients.
In this work, we examine the effectiveness of an uncertainty quantification framework known as Evidential Deep Learning applied in the context of biomedical image segmentation. This class of models involves assigning Dirichlet distributions as priors for segmentation labels, and enables a few distinct definitions of model uncertainties. Using the cardiac and prostate MRI images available in the Medical Segmentation Decathlon for validation, we found that Evidential Deep Learning models with U-Net backbones generally yielded superior correlations between prediction errors and uncertainties relative to the conventional baseline equipped with Shannon entropy measure, Monte-Carlo Dropout and Deep Ensemble methods. We also examined these models' effectiveness in active learning, finding that relative to the standard Shannon entropy-based sampling, they yielded higher point-biserial uncertainty-error correlations while attaining similar performances in Dice-Sorensen coefficients. These superior features of EDL models render them well-suited for segmentation tasks that warrant a critical sensitivity in detecting large model errors.