Table 2 Ablation study on FFN.
From: Feature enhanced cascading attention network for lightweight image super-resolution
FFN | Params | FLOPs | PSNR (dB) | ||
|---|---|---|---|---|---|
MLP | SGFF | [K] | [G] | Set5 | B100 |
✕ | ✓ | 121 | 6.68 | 32.08 | 27.57 |
✓ | ✕ | 128 | 7.08 | 32.07 | 27.55 |
✕ | ✕ | 92 | 5.05 | 31.98 | 27.50 |
From: Feature enhanced cascading attention network for lightweight image super-resolution
FFN | Params | FLOPs | PSNR (dB) | ||
|---|---|---|---|---|---|
MLP | SGFF | [K] | [G] | Set5 | B100 |
✕ | ✓ | 121 | 6.68 | 32.08 | 27.57 |
✓ | ✕ | 128 | 7.08 | 32.07 | 27.55 |
✕ | ✕ | 92 | 5.05 | 31.98 | 27.50 |