Table 6 Effect of MSSA and MSCA.
From: Federated learning with integrated attention multiscale model for brain tumor segmentation
Size of kernel | #Params (M) | #GFLOPs | Dice (%) |
|---|---|---|---|
1 × 1 | 21.52 | 31.81 | 92.32 |
3 × 3 | 21.92 | 32.58 | 93.02 |
5 × 5 | 21.95 | 32.65 | 92.83 |
7 × 7 | 21.99 | 32.75 | 92.56 |
All scales | 21.89 | 32.54 | 93.28 |
