Table 9 Performance of each component (only decoder) on the ISIC dataset (ISIC17 on the top, and ISIC18 on the bottom).
From: VMKLA-UNet: vision Mamba with KAN linear attention U-Net
Model | SS2D | Linear | KAN Linear | CSA | mIoU (%) | DSC (%) | Acc (%) |
|---|---|---|---|---|---|---|---|
Baseline | √ | 80.23 | 89.03 | 96.29 | |||
Only-CSA | √ | 81.88 | 90.12 | 96.68 | |||
MLLA | √ | 78.87 | 88.40 | 96.17 | |||
MKLA | √ | 81.12 | 89.16 | 95.87 | |||
MLCSA | √ | √ | 83.37 | 90.23 | 96.83 | ||
Ours | √ | √ | 84.51 | 91.60 | 97.39 | ||
Baseline | √ | 81.35 | 89.71 | 94.91 | |||
Only-CSA | √ | 82.76 | 89.91 | 95.56 | |||
MLLA | √ | 81.33 | 90.23 | 94.99 | |||
MKLA | √ | 82.45 | 90.32 | 95.46 | |||
MLCSA | √ | √ | 83.19 | 91.23 | 96.03 | ||
Ours | √ | √ | 84.16 | 91.40 | 96.14 |
