Table 3 Performance of the models over the test dataset.
Models | Classes | IoU (%) | Recall (%) | Precision (%) | mIoU (%) | mPA (%) | Accuracy (%) |
|---|---|---|---|---|---|---|---|
UNet | W | 92.79 | 96.82 | 95.71 | 82.84 | 89.58 | 93.96 |
B | 72.88 | 82.33 | 86.40 | ||||
DeepLabv3+ | W | 91.57 | 96.25 | 94.95 | 80.12 | 87.71 | 92.88 |
B | 68.67 | 79.16 | 83.82 | ||||
HRNet | W | 92.45 | 96.46 | 95.33 | 82.93 | 89.62 | 92.95 |
B | 72.19 | 83.16 | 85.82 | ||||
SegFormer | W | 92.50 | 96.80 | 96.29 | 83.13 | 90.23 | 93.69 |
B | 73.77 | 83.66 | 86.78 | ||||
SegFormer (RGB + TIFF) | W | 92.87 | 96.76 | 95.86 | 83.08 | 89.86 | 94.04 |
B | 73.28 | 82.97 | 86.26 | ||||
Tiff-SegFormer | W | 93.63 | 96.83 | 96.59 | 84.94 | 91.46 | 94.71 |
B | 76.25 | 86.09 | 86.96 |
