Table 5.

Results of student with simple augmentations model (ResNet34 backbone) depending on train set composition on PH2 dataset.

	mIoU	Dice	Precision	Recall
teacher	0.8574	0.9192	0.8502	0.9760
m = 1	0.8554	0.9181	0.8459	0.9782
m = 2	0.8575	0.9200	0.8475	0.9791
m = 4	0.8603	0.9222	0.8484	0.9811
m = 8	0.8537	0.9135	0.8380	0.9825