Table 2.

Results of student with simple augmentations model (ResNet18 backbone) depending on train set composition on ISIC 2018 dataset.

	mIoU	Dice	Precision	Recall
teacher	0.8659	0.9194	0.8781	0.9308
m = 1	0.8789	0.9355	0.9030	0.9303
m = 2	0.8713	0.9215	0.8690	0.9482
m = 4	0.8657	0.9245	0.8760	0.9382
m = 8	0.8714	0.9292	0.8808	0.9396