Skip to main content
. 2022 Dec 6;16:1043569. doi: 10.3389/fnhum.2022.1043569

Table 3.

Comparison of segmentation methods using domain adaptation.

Methods Drishti-GS dataset RIM-ONE-r3 dataset
DSC cup DSC disc δ DSC cup DSC disc δ
BEAL (Wang et al., 2019a) 0.862 0.961 - 0.810 0.898 -
DoFE (Wang S. et al., 2020) 0.835 0.955 - 0.800 0.893 -
CADA (Liu et al., 2022) 0.840 0.890 0.111 0.640 0.766 0.087
CFEA (Liu et al., 2019) 0.827 0.887 0.113 0.635 0.751 0.095
AdaptSegNet (Tsai et al., 2018) 0.826 0.881 0.118 0.627 0.737 0.102
DAE (Karani et al., 2021) 0.831 0.940 - 0.790 0.891 -
SRDA (Bateson et al., 2020) 0.807 0.962 - 0.776 0.894 -
Ours 0.857 0.962 0.083 0.791 0.898 0.082

The best result of the column is shown in bold.