Is Segmentation Uncertainty Useful?

Probabilistic image segmentation encodes varying prediction confidence and inherent ambiguity in the problem. While different probabilistic models are designed to capture aspects of uncertainty ambiguity, these modelling differences rarely discussed context applications uncertainty. We consider two common use cases uncertainty, namely assessment quality active learning. four established strategies for segmentation, discuss their capabilities, investigate performance tasks. find that all both tasks, returned correlates positively with error, but does not prove be useful