Table 13 Quantitative comparison of ablation experiments on loss functions using the LOL dataset.
From: A multi scale spatial attention based zero shot learning framework for low light image enhancement
Method | PSNR | SSIM | FSIM |
|---|---|---|---|
w/o \(L_{tv}\) | 11.88 | 0.661 | 0.751 |
w/o \(L_{spa}\) | 14.36 | 0.768 | 0.858 |
w/o \(L_{color}\) | 4.67 | 0.506 | 0.596 |
w/o \(L_{bri}\) | 8.89 | 0.456 | 0.546 |
w/o \(L_{sem}\) | 15.98 | 0.769 | 0.859 |
w/o \(L_{NR}\) | 14.95 | 0.734 | 0.824 |
Ours | 18.39 | 0.850 | 0.940 |
