Table 1 Comparison of segmentation performance of different residual networks
Algorithm | IoU (%) | Dice (%) | Accuracy (%) | mIoU (%) | mDice (%) |
|---|---|---|---|---|---|
FCN | 61.72 ± 5.24 | 76.33 ± 4.22 | 67.86 ± 7.32 | 80.88 ± 2.63 | 88.15 ± 2.11 |
Deeplabv3 | 69.04 ± 5.87 | 81.66 ± 6.11 | 77.32 ± 6.57 | 84.49 ± 2.94 | 90.83 ± 2.23 |
Deeplabv3plus | 72.87 ± 4.49 | 84.31 ± 3.07 | 78.52 ± 4.33 | 86.42 ± 2.25 | 92.14 ± 1.54 |
PlaqueNet | 87.37 ± 10.38 | 93.26 ± 8.36 | 93.12 ± 10.66 | 93.68 ± 5.20 | 96.63 ± 4.19 |