Table 2.
Statistical segmentation quality measurement by using COVID-SVR Data of the proposed against active contour [49], GraphCut [50], ResNet [51], Gated-UNet [60], Dense UNet [61], Robust Threshold [52], [53], and Tunable Weka [54]. The best results are shown in red.
Segmentation methods | Quality measurements |
||
---|---|---|---|
Dice (SDC) ↑ | Jaccard ↑ | Hausdorff ↓ | |
Proposed | 0.49 | 0.46 | 5.62 |
Active Contour [49] | 0.30 | 0.32 | 17.77 |
GraphCut [50] | 0.29 | 0.29 | 18.64 |
ResNet [51] | 0.33 | 0.31 | 27 |
Robust Threshold [52], [53] | 0.28 | 0.34 | 15.81 |
Gated-UNet [60] | 0.44 | – | – |
Dense-UNet [61] | 0.41 | – | – |
Tunable Weka [54] | 0.51 | 0.43 | 9.68 |