Table 3.
Accuracy metrics comparison for different segmentation methods.
| Methods | IoU | Accuracy (%) | Precision (%) | Recall (%) | F1-Score |
|---|---|---|---|---|---|
| Unet (resnet34) | 78.44 | 88.65 | 84.69 | 91.90 | 87.06 |
| DeepLabV3+(resnet34) | 81.83 | 90.62 | 86.65 | 92.40 | 90.31 |
| FPN (resnet34) | 79.84 | 89.79 | 93.13 | 84.88 | 88.59 |
| PSPnet (resnet34) | 83.18 | 92.25 | 91.64 | 91.52 | 89.49 |
| DSCA-PSPNet | 87.58 | 92.34 | 93.80 | 93.21 | 92.38 |
Bold value is the highest value.
Underline value is the second highest value.