Method | PA | mIoU | Dice | FPN | ||||
D1 | D2 | D1 | D2 | D1 | D2 | D1 | D2 | |
FCN [7] | 0.899 | 0.875 | 0.905 | 0.869 | 0.812 | 0.844 | 0.311 | 0.315 |
DeepLabv3+ [8] | 0.931 | 0.890 | 0.898 | 0.901 | 0.902 | 0.917 | 0.210 | 0.291 |
SegNet [9] | 0.929 | 0.911 | 0.906 | 0.899 | 0.932 | 0.890 | 0.294 | 0.332 |
Mask RCNN [10] | 0.891 | 0.879 | 0.884 | 0.903 | 0.878 | 0.889 | 0.295 | 0.344 |
nn U-Net [11] | 0.877 | 0.855 | 0.781 | 0.873 | 0.924 | 0.927 | 0.209 | 0.254 |
BCDU-Net [12] | 0.879 | 0.907 | 0.859 | 0.891 | 0.891 | 0.866 | 0.195 | 0.265 |
Medical Transformer [13] | 0.904 | 0.890 | 0.901 | 0.900 | 0.918 | 0.879 | 0.214 | 0.277 |
TransUNe [14] | 0.925 | 0.917 | 0.863 | 0.895 | 0.894 | 0.905 | 0.203 | 0.253 |
CaraNet [15] | 0.911 | 0.869 | 0.892 | 0.897 | 0.909 | 0.890 | 0.218 | 0.217 |
Ours | 0.948 | 0.937 | 0.913 | 0.935 | 0.934 | 0.915 | 0.187 | 0.197 |