Image Denoising on PolyU SIDD pre-trained real-world (test)
39.28PSNRLAN
Evaluation results
| Method | Links | ||
|---|---|---|---|
| LANBackbone=Restormer, Iteration=20, Loss=ZS-N2N2024.12 | 39.28 | 0.969 | |
| LANBackbone=Uformer, Iteration=20, Loss=ZS-N2N2024.12 | 39.2 | 0.968 | |
| LANBackbone=Restormer, Iteration=20, Loss=Nbr2Nbr2024.12 | 39.17 | 0.968 | |
| LANBackbone=Uformer, Iteration=20, Loss=Nbr2Nbr2024.12 | 39.1 | 0.967 | |
| pretrainedBackbone=Restormer, Iteration=0, Adaptation=None2024.12 | 39.03 | 0.966 | |
| pretrainedBackbone=Uformer, Iteration=0, Adaptation=None2024.12 | 38.93 | 0.965 | |
| LANBackbone=DnCNN, Iteration=20, Loss=Nbr2Nbr2024.12 | 38.31 | 0.956 | |
| LANBackbone=DnCNN, Iteration=10, Loss=ZS-N2N2024.12 | 38.29 | 0.955 | |
| LANBackbone=DnCNN, Iteration=20, Loss=ZS-N2N2024.12 | 38.29 | 0.955 | |
| meta-learningBackbone=DnCNN, Iteration=10, Loss=Nbr2Nbr2024.12 | 38.25 | 0.955 | |
| meta-learningBackbone=DnCNN, Iteration=5, Loss=ZS-N2N2024.12 | 38.23 | 0.955 | |
| meta-learningBackbone=DnCNN, Iteration=5, Loss=Nbr2Nbr2024.12 | 38.23 | 0.956 | |
| meta-learningBackbone=DnCNN, Iteration=10, Loss=ZS-N2N2024.12 | 38.23 | 0.955 | |
| LANBackbone=DnCNN, Iteration=5, Loss=ZS-N2N2024.12 | 38.22 | 0.954 | |
| LANBackbone=DnCNN, Iteration=10, Loss=Nbr2Nbr2024.12 | 38.22 | 0.954 | |
| meta-learningBackbone=DnCNN, Iteration=20, Loss=Nbr2Nbr2024.12 | 38.2 | 0.954 | |
| meta-learningBackbone=DnCNN, Iteration=20, Loss=ZS-N2N2024.12 | 38.17 | 0.953 | |
| LANBackbone=DnCNN, Iteration=5, Loss=Nbr2Nbr2024.12 | 38.16 | 0.953 | |
| last-layerBackbone=DnCNN, Iteration=10, Loss=Nbr2Nbr2024.12 | 38.14 | 0.952 | |
| last-layerBackbone=DnCNN, Iteration=20, Loss=Nbr2Nbr2024.12 | 38.14 | 0.952 | |
| last-layerBackbone=DnCNN, Iteration=5, Loss=Nbr2Nbr2024.12 | 38.13 | 0.952 | |
| last-layerBackbone=DnCNN, Iteration=10, Loss=ZS-N2N2024.12 | 38.13 | 0.952 | |
| last-layerBackbone=DnCNN, Iteration=20, Loss=ZS-N2N2024.12 | 38.13 | 0.952 | |
| last-layerBackbone=DnCNN, Iteration=5, Loss=ZS-N2N2024.12 | 38.12 | 0.952 | |
| pretrainedBackbone=DnCNN, Iteration=0, Adaptation=None2024.12 | 38.1 | 0.952 | |
| full-trainableBackbone=DnCNN, Iteration=5, Loss=Nbr2Nbr2024.12 | 38.08 | 0.951 | |
| full-trainableBackbone=DnCNN, Iteration=5, Loss=ZS-N2N2024.12 | 38.07 | 0.951 | |
| full-trainableBackbone=DnCNN, Iteration=10, Loss=Nbr2Nbr2024.12 | 38.06 | 0.951 | |
| full-trainableBackbone=DnCNN, Iteration=10, Loss=ZS-N2N2024.12 | 38.04 | 0.95 | |
| full-trainableBackbone=DnCNN, Iteration=20, Loss=Nbr2Nbr2024.12 | 38.02 | 0.949 | |
| full-trainableBackbone=DnCNN, Iteration=20, Loss=ZS-N2N2024.12 | 37.99 | 0.949 | |
| first-layerBackbone=DnCNN, Iteration=5, Loss=Nbr2Nbr2024.12 | 37.95 | 0.948 | |
| first-layerBackbone=DnCNN, Iteration=5, Loss=ZS-N2N2024.12 | 37.93 | 0.948 | |
| first-layerBackbone=DnCNN, Iteration=10, Loss=Nbr2Nbr2024.12 | 37.83 | 0.945 | |
| first-layerBackbone=DnCNN, Iteration=10, Loss=ZS-N2N2024.12 | 37.76 | 0.943 | |
| first-layerBackbone=DnCNN, Iteration=20, Loss=Nbr2Nbr2024.12 | 37.67 | 0.941 | |
| first-layerBackbone=DnCNN, Iteration=20, Loss=ZS-N2N2024.12 | 37.47 | 0.935 |