From: Facial image super-resolution guided by adaptive geometric features
Network architecture | ||
---|---|---|
The main SR block | The depth estimation block | The modulation block |
Conv: [3×3,3,64] | Conv: [3×3,3,64] | Conv: [3×3,1,64] |
Residual: [3×3,64,64]×8 | Downsample: [3×3,3,64]×3 | |
Feature addition | Residual: [3×3,64,64]×5 | |
Feature multiplication | Upsample: [3×3,64,64] | |
Residual: [3×3,64,64]×8 | Feature concatenation | |
Feature addition | Upsample: [3×3,64,64] | Conv: [3×3,64,64]×3 |
Conv: [3×3,64,64] | Feature concatenation | |
Upsample: [3×3,64,64]×2 | Upsample: [3×3,64,64] | |
Conv: [3×3,64,3] | Feature concatenation | |
Output image | Conv: [3×3,64,1] |