Table 3.
Segmentation performance evaluation on Mask-RCNN and U-Net.
| Methods | Mask-RCNN | U-Net | ||
|---|---|---|---|---|
| DSC (%) | HD | DSC (%) | HD | |
| Baseline | 75.70 ± 3.792 | 121.75 ± 44.292 | 64.87 ± 2.066 | 159.48 ± 15.185 |
| RefMeth_1 | 75.98 ± 3.271 | 110.84 ± 14.223 | 65.28 ± 2.245 | 179.32 ± 6.2712 |
| RefMeth_2 | 77.34 ± 1.774 | 79.803 ± 8.744 | 65.55 ± 1.757 | 164.46 ± 7.9453 |
| RefMeth_3 | 78.88 ± 3.000 | 73.920 ± 13.632 | 63.82 ± 3.299 | 169.90 ± 9.9893 |
| RefMeth_4 | 80.52 ± 2.718 | 66.865 ± 14.626 | 66.85 ± 3.422 | 171.41 ± 11.353 |
| PropMeth_1 | 85.36 ± 1.156 | 43.523 ± 2.7343 | 77.30 ± 2.900 | 66.869 ± 4.6270 |
| PropMeth_2 | 85.16 ± 1.459 | 44.410 ± 6.5669 | 78.61 ± 2.616 | 59.627 ± 8.068 |
| PropMeth_3 | 85.44 ± 1.062 | 44.641 ± 4.7393 | 79.08 ± 2.654 | 66.070 ± 2.1231 |