Table 7 Performance of encoders with different depths on the ISIC dataset.
From: VMKLA-UNet: vision Mamba with KAN linear attention U-Net
Model | Depth | Dataset | mIoU (%) | DSC (%) | Acc (%) |
|---|---|---|---|---|---|
VMKLA-T | [2, 2, 2, 2] | ISIC17 | 83.34 | 90.91 | 97.21 |
ISIC18 | 82.98 | 90.70 | 95.75 | ||
VMKLA-S | [2, 2, 9, 2] | ISIC17 | 84.37 | 91.52 | 97.39 |
ISIC18 | 85.10 | 91.95 | 96.37 | ||
VMKLA-B | [2, 2, 27, 2] | ISIC17 | 84.51 | 91.60 | 97.39 |
ISIC18 | 84.16 | 91.40 | 96.14 |
