Fig. 4
From: Label-set impact on deep learning-based prostate segmentation on MRI
Performance of model 4, model 5 and manual segmentations on set A and set B. Boxplots displaying the Dice similarity coefficient (DSC) (a and d), the 95th percentile of the Hausdorff distance (HD95; log applied to improve visualization) (b and e), and relative volume difference for whole prostate (RVD (WP)) (c and f) comparing agreement between the masks of the test subset of the PROSTATEx dataset generated by model 4 and model 5 with set A (a–c), and set B (d–f) as reference, in addition to the similarity between set A and set B (human). ns: p ≥ 0.05, * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001