Table 1.

Comparison of segmentation results of the unsupervised domain adaptation methods from the top-ranked methods of the CrossMoDA 2021, 2022 challenge. The paired t-test was conducted to assess the statistical significance of our results. $(\ast p<0.05,\ast \ast p<0.0001)$.

Method	Dice score ($\uparrow$)			ASSD ($\downarrow$)
	VS	Cochlea	Mean	VS	Cochlea
COSMOS13	0.6104 (±0.3065)**	0.8184 (±0.0257)**	0.7144 (±0.1480)	3.8170 (±5.0257)**	0.2293 (±0.1633)**
Fgh36521	0.8178 (±0.0803)**	0.8433 (±0.0293)**	0.8306 (±0.0420)	0.6673 (±0.2713)**	0.2053 (±0.1489)**
MSF-Net43	0.8493 (±0.0683)	0.8294 (±0.0268)**	0.8394 (±0.0368)	0.5202 (±0.2288)**	0.2454 (±0.2102)**
PAST14	0.8473 (±0.0633)	0.8547 (±0.0283)**	0.8511 (±0.0322)	0.5513 (±0.3026)**	0.1874 (±0.1478)**
Multi-view17	0.8520 (±0.0889)	0.8488 (±0.0235)**	0.8504 (±0.0466)	0.4748 (±0.2072)	0.1992 (±0.1524)**
TBA34	0.8682 (±0.0601)*	0.8506 (±0.0294)**	0.8594 (±0.0347)	0.4302 (±0.1780)	0.1892 (±0.1457)**
Ours	0.8550 (±0.0731)	0.8750 (±0.0217)	0.8650 (±0.0370)	0.4643 (±0.2000)	0.1553 (±0.1448)