Table 2 Quantitative comparison of the proposed model and other deep learning models.
Methods | Precision | Recall | IoU | F1 | ||||
|---|---|---|---|---|---|---|---|---|
IS | PS | IS | PS | IS | PS | IS | PS | |
CA-DeeplabV3+ | 93.19% | 88.37% | 92.95% | 88.75% | 87.04% | 79.47% | 93.07% | 88.56% |
DeeplabV3+ | 90.87% | 88.47% | 93.35% | 84.46% | 85.35% | 76.08% | 92.10% | 86.41% |
UNet | 88.78% | 85.07% | 91.51% | 80.69% | 83.74% | 74.78% | 90.13% | 82.82% |
PSPNet | 91.20% | 88.78% | 92.57% | 84.93% | 84.24% | 75.68% | 92.37% | 86.81% |
