Table 1. Comparison with state-of-the-art semi-supervised methods on the DME dataset and Metrics reported the mean ± standard results with three random seeds.

Method	Training		Metrics
	Labeled	Unlabeled	DSC(%)↑	Jaccard(%)↑	95HD(voxel)↓	ASD(voxel)↓
V-Net [29]	4(10%)	0	75.90 ± 6.13	61.36 ± 6.41	7.81 ± 5.51	2.87 ± 1.89
V-Net [29]	10(25%)	0	83.13 ± 4.42	70.79 ± 5.29	5.34 ± 3.44	1.54 ± 1.21
V-Net [29]	40(All)	0	88.23 ± 0.87	78.65 ± 1.24	1.86 ± 0.32	0.82 ± 0.43
UAMT [22]	4(10%)	36(90%)	82.28 ± 3.92	70.08 ± 5.67	5.69 ± 3.96	2.07 ± 1.62
CAML [37]			83.54 ± 3.31	71.77 ± 4.13	4.65 ± 1.72	1.53 ± 1.13
MCNet [35]			86.06 ± 1.28	72.10 ± 1.79	3.21 ± 1.79	0.89 ± 0.26
URPC [36]			78.21 ± 3.56	64.36 ± 4.80	5.90 ± 0.80	1.17 ± 0.77
Ours			87.50 ± 0.91	77.79 ± 1.43	2.77 ± 1.35	0.79 ± 0.38
UAMT [22]	10(25%)	30(75%)	87.42 ± 1.24	77.67 ± 1.95	2.21 ± 0.79	0.80 ± 0.39
CAML [37]			85.12 ± 0.63	74.10 ± 0.95	2.81 ± 0.36	0.80 ± 0.36
MCNet [35]			88.88 ± 0.29	79.99 ± 0.41	1.71 ± 0.29	0.67 ± 0.29
URPC [36]			84.70 ± 0.83	73.47 ± 1.24	3.50 ± 0.50	0.73 ± 0.26
Ours			91.04 ± 0.24	81.93 ± 0.24	1.38 ± 0.29	0.56 ± 0.27