Image Denoising on Monte Carlo rendered image 128 spp
30.62PSNROurs
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| OursTraining noise protocol=Gaussian noise, σ = 15, Training strategy=masked training2023.03 | 30.62 | 85 | 0.2254 | |
| RIDNetTraining noise protocol=Gaussian noise, σ = 152023.03 | 29.96 | 79.21 | 0.2548 | |
| DnCNNTraining noise protocol=Gaussian noise, σ = 152023.03 | 29.94 | 78.83 | 0.2671 | |
| RNANTraining noise protocol=Gaussian noise, σ = 152023.03 | 29.86 | 78.25 | 0.2702 | |
| baselineTraining noise protocol=Gaussian noise, σ = 152023.03 | 29.68 | 77.38 | 0.2851 | |
| SwinIRTraining noise protocol=Gaussian noise, σ = 152023.03 | 29.32 | 76.27 | 0.2943 | |
| DropoutTraining noise protocol=Gaussian noise, σ = 152023.03 | 28.85 | 77.53 | 0.2941 | |
| RestormerTraining noise protocol=Gaussian noise, σ = 152023.03 | 24.98 | 65.98 | 0.4575 |