Fig. 2: Segmentation performance of the two approximate Bayesian methods. | Communications Medicine

Fig. 2: Segmentation performance of the two approximate Bayesian methods.

From: Application of simultaneous uncertainty quantification and segmentation for oropharyngeal cancer use-case with Bayesian deep learning

Fig. 2

Violinplot of Dice similarity coefficient (DSC), mean surface distance (MSD), and Hausdorff distance at 95% (95HD) performance on the external dataset (N = 67). The dotted lines mark the quartiles. Statistical significance is measured using the Wilcoxon signed-rank test.

Back to article page