Table 1 Total variance (TV) loss for different layers of CNN features.
From: Low-light image enhancement via adaptive frequency decomposition network
Features | T1 layer | T2 layer | T3 layer | T4 layer |
|---|---|---|---|---|
U-Net | 0.3913 | 0.2238 | 0.2355 | 0.1493 |
U-Net with AFD module | 0.3946 | 0.4526 | 0.0665 | 0.0863 |
