Table 1 Evaluation indicators ± standard deviation of all competing methods in the Synapse multi-organ segmentation dataset.
Methods | DICE (%) | IOU (%) | RAVD (%) | ASSD | MSSD |
|---|---|---|---|---|---|
FCN | 62.61 ± 0.56 | 49.67 ± 0.46 | − 17.67 ± 0.48 | 7.98 ± 0.48 | 43.11 ± 4.2 |
U-Net | 75.73 ± 0.45 | 64.52 ± 0.52 | − 8.07 ± 0.39 | 5.83 ± 0.39 | 32.50 ± 3.8 |
ResU-Net | 77.31 ± 0.33 | 66.40 ± 0.33 | − 10.85 ± 0.22 | 5.49 ± 0.52 | 31.00 ± 2.5 |
Attention U-Net | 77.57 ± 0.42 | 66.44 ± 0.28 | − 9.91 ± 0.45 | 5.29 ± 0.37 | 31.35 ± 2.9 |
U-Net++ | 79.33 ± 0.33 | 67.53 ± 0.36 | − 8.68 ± 0.12 | 4.74 ± 0.28 | 26.56 ± 5.1 |
ResU-Net++ | 76.61 ± 0.24 | 65.00 ± 0.28 | 3.21 ± 0.36 | 4.76 ± 0.42 | 25.52 ± 2.7 |
TransformU-Net | 79.52 ± 0.14 | 66.48 ± 0.19 | 1.63 ± 0.25 | 4.33 ± 0.37 | 23.51 ± 3.2 |
SwimU-Net | 76.61 ± 0.31 | 65.00 ± 0.20 | 3.21 ± 0.21 | 4.76 ± 0.22 | 25.52 ± 2.9 |
HiFormer | 76.92 ± 0.21 | 65.4 ± 0.36 | − 5.56 ± 0.15 | 4.53 ± 0.38 | 26.81 ± 2.8 |
D2HU-Net | 80.39 ± 0.19 | 68.32 ± 0.13 | 1.54 ± 0.09 | 4.03 ± 0.36 | 21.49 ± 2.4 |
