Table 3. Comparison with other methods that may improve the annotation efficiency in OCT segmentation. The data format in the table is the mean (standard deviation).

Methods		SRF		PED		Choroid
		10%	100%	10%	100%	10%	100%
Transfer learning [29]		0.637 (0.278)	0.727 (0.239)	0.381 (0.212)	0.525 (0.254)	0.875 (0.087)	0.890 (0.076)
Semi-supervised learning-1 [48]		0.784 (0.220)	0.796 (0.198)	0.587 (0.244)	0.622 (0.255)	0.871 (0.076)	0.882 (0.095)
Semi-supervised learning-2 [49]		0.737 (0.267)	0.798 (0.193)	0.536 (0.229)	0.672 (0.212)	0.874 (0.081)	0.888 (0.084)
Ours		0.795 (0.186)	0.824 (0.148)	0.676 (0.209)	0.724 (0.196)	0.879 (0.077)	0.895 (0.073)