Multispectral Demosaicing on HyKo
34.81PSNRPEFD
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| PEFD2026.03 | 34.81 | 93.8 | 0.064 | |
| GaussianKernel Size=9x92026.03 | 32.57 | 91.5 | 0.044 | |
| Garcia-Barajas et al.Architecture=DnCNN [53], Loss Function=rotate-EI loss [7]2026.03 | 32.57 | 91.5 | 0.045 | |
| PPID2026.03 | 31.46 | 90.1 | 0.052 | |
| Feng et al.Architecture=SDNet, Loss Function=shift-EI loss [6]2026.03 | 30.95 | 90 | 0.052 | |
| Liu et al.Loss Function=limited-pixel shift-equivariant loss2026.03 | 30.56 | 85.8 | 0.118 | |
| BilinearKernel Size=5x52026.03 | 30.55 | 90 | 0.064 | |
| Weighted-Bilin.Kernel Size=7x72026.03 | 30.38 | 89.6 | 0.055 | |
| Li et al.Base Model=EDSR [34]2026.03 | 29.15 | 88.3 | 0.059 | |
| DIPArchitecture=16-channel convolutional decoder network2026.03 | 28.43 | 88.8 | 0.168 | |
| TVOptimization=proximal gradient descent2026.03 | 23.08 | 74.4 | 0.045 | |
| MCFinetuning Strategy=measurement consistency2026.03 | 21.74 | 55.6 | 0.169 | |
| Mosaiced2026.03 | 11.33 | 7.2 | 1.33 |