Calibrating the Dice loss to handle neural network overconfidence for biomedical image segmentation

Open in new window