Noise Synthesis on SIDD (test)
0.008DKLNoise Flow
Evaluation results
| Method | Links | |
|---|---|---|
| Noise Flow2019.08 | 0.008 | |
| Contrastive Noise Estimationcamera-specific training data=false2022.04 | 0.01 | |
| CANGAN2022.04 | 0.0178 | |
| Noiseflow2022.04 | 0.0327 | |
| Contrastive Noise Estimationcamera-specific training data=false2022.04 | 0.0385 | |
| CANGAN2022.04 | 0.0432 | |
| P-G2022.04 | 0.0514 | |
| Cam. NLFmodel_type=heteroscedastic Gaussian, source=camera-calibrated2019.08 | 0.052 | |
| Noiseflow2022.04 | 0.0617 | |
| P-G2022.04 | 0.1023 | |
| Gaussianmodel_type=homoscedastic Gaussian, estimation=Maximum Likelihood Estimate (MLE)2019.08 | 0.394 | |
| AWGN2022.04 | 0.4793 | |
| AWGN2022.04 | 0.8367 |