Gaussian Color Image Denoising on Urban100 (test)
30.46PSNR (sigma=50)GRL-B
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| GRL-BParams [M]=19.812023.03 | 30.46 | 35.54 | 33.35 | — | |
| DMID-dTraining strategy=Noise-blind (various noise levels)2023.07 | 30.28 | 35.26 | 33.11 | — | |
| Restormer-LocalTraining strategy=Separate models2021.12 | 30.17 | 35.21 | 33.06 | — | |
| Restormer-LocalTraining strategy=Single model2021.12 | 30.16 | 35.14 | 33.01 | — | |
| EDT-BParams [M]=11.482023.03 | 30.16 | 35.22 | 33.07 | — | |
| GRL-SParams [M]=3.122023.03 | 30.09 | 35.24 | 33.07 | — | |
| RestormerTraining strategy=single model2021.11 | 30.02 | 35.06 | 32.91 | — | |
| RestormerTraining strategy=separate models2021.11 | 30.02 | 35.13 | 32.96 | — | |
| RestormerTraining strategy=Single model2021.12 | 30.02 | 35.06 | 32.91 | — | |
| RestormerTraining strategy=Separate models2021.12 | 30.02 | 35.13 | 32.96 | — | |
| RestormerParams [M]=26.132023.03 | 30.02 | 35.13 | 32.96 | — | |
| RestormerTraining strategy=Noise-specific (separate models)2023.07 | 30.02 | 35.13 | 32.96 | — | |
| RestormerTraining strategy=Noise-blind (various noise levels)2023.07 | 30.02 | 35.06 | 32.91 | — | |
| SwinIRTraining strategy=separate models2021.11 | 29.82 | 35.13 | 32.9 | — | |
| SwinIRTraining strategy=Separate models2021.12 | 29.82 | 35.13 | 32.9 | — | |
| SwinIRParams [M]=11.752023.03 | 29.82 | 35.13 | 32.9 | — | |
| SwinIRTraining strategy=Noise-specific (separate models)2023.07 | 29.82 | 35.13 | 32.9 | — | |
| GRL-TParams [M]=0.882023.03 | 29.78 | 35.08 | 32.84 | — | |
| IPTTraining strategy=separate models2021.11 | 29.71 | — | — | — | |
| IPTTraining strategy=Separate models2021.12 | 29.71 | — | — | — | |
| IPTParams [M]=115.332023.03 | 29.71 | — | — | — | |
| IPTTraining strategy=Noise-specific (separate models)2023.07 | 29.71 | — | — | — | |
| DRUNetTraining strategy=single model2021.11 | 29.61 | 34.81 | 32.6 | — | |
| DRUNetTraining strategy=Single model2021.12 | 29.61 | 34.81 | 32.6 | — | |
| DRUNetParams [M]=32.642023.03 | 29.61 | 34.81 | 32.6 | — | |
| DRUNetTraining strategy=Noise-blind (various noise levels)2023.07 | 29.61 | 34.81 | 32.6 | — | |
| RDNTraining strategy=separate models2021.11 | 29.38 | — | — | — | |
| RDNTraining strategy=Separate models2021.12 | 29.38 | — | — | — | |
| RDNTraining strategy=Noise-specific (separate models)2023.07 | 29.38 | — | — | — | |
| RNANTraining strategy=separate models2021.11 | 29.08 | — | — | — | |
| RNANTraining strategy=Separate models2021.12 | 29.08 | — | — | — | |
| RNANParams [M]=8.962023.03 | 29.08 | — | — | — | |
| RNANTraining strategy=Noise-specific (separate models)2023.07 | 29.08 | — | — | — | |
| RPCNNTraining strategy=separate models2021.11 | 28.62 | — | 31.81 | — | |
| RPCNNTraining strategy=Separate models2021.12 | 28.62 | — | 31.81 | — | |
| BRDNetTraining strategy=separate models2021.11 | 28.56 | 34.42 | 31.99 | — | |
| BRDNetTraining strategy=Separate models2021.12 | 28.56 | 34.42 | 31.99 | — | |
| BRDNetTraining strategy=Noise-specific (separate models)2023.07 | 28.56 | 34.42 | 31.99 | — | |
| KBNet_sMACs=69G2023.03 | 28.33 | 33.77 | 31.45 | — | |
| RestormerMACs=141G2023.03 | 28.29 | 33.79 | 31.46 | — | |
| FFDNetTraining strategy=single model2021.11 | 28.05 | 33.83 | 31.4 | — | |
| FFDNetTraining strategy=Single model2021.12 | 28.05 | 33.83 | 31.4 | — | |
| FFDNetTraining strategy=Noise-blind (various noise levels)2023.07 | 28.05 | 33.83 | 31.4 | — | |
| SwinIRMACs=759G2023.03 | 27.98 | 33.7 | 31.3 | — | |
| DAGLMACs=256G2023.03 | 27.97 | 33.79 | 31.39 | — | |
| DRUNetMACs=144G2023.03 | 27.96 | 33.41 | 31.11 | — | |
| IRCNNTraining strategy=single model2021.11 | 27.7 | 33.78 | 31.2 | — | |
| IRCNNTraining strategy=Single model2021.12 | 27.7 | 33.78 | 31.2 | — | |
| RNANMACs=496G2023.03 | 27.65 | — | — | — | |
| DnCNNTraining strategy=single model2021.11 | 27.59 | 32.98 | 30.81 | — | |
| DnCNNTraining strategy=Single model2021.12 | 27.59 | 32.98 | 30.81 | — | |
| DnCNNParams [M]=0.562023.03 | 27.59 | 32.98 | 30.81 | — | |
| DnCNNTraining strategy=Noise-blind (various noise levels)2023.07 | 27.59 | 32.98 | 30.81 | — | |
| DeamNetMACs=146G2023.03 | 27.53 | 33.37 | 30.85 | — | |
| NLRN2023.03 | 27.49 | 33.45 | 30.94 | — | |
| MWCNN2023.03 | 27.42 | 33.17 | 30.66 | — | |
| FOCNet2023.03 | 27.4 | 33.15 | 30.64 | — | |
| FFDNet2023.03 | 26.5 | 32.4 | 29.9 | — | |
| DnCNNMACs=37G2023.03 | 26.35 | 32.28 | 29.8 | — | |
| IRCNN2023.03 | 26.22 | 32.46 | 29.8 | — | |
| CBM3DTraining strategy=Noise-blind (various noise levels)2023.07 | 25.95 | 32.35 | 29.7 | — | |
| DnCNNnoise level (σ)=152024.02 | — | — | — | 32.98 | |
| DnCNNnoise level (σ)=252024.02 | — | — | — | 30.81 | |
| DnCNNnoise level (σ)=502024.02 | — | — | — | 27.59 | |
| DnCNNσ=152024.11 | — | — | — | 32.98 | |
| DRUNetnoise level (σ)=152024.02 | — | — | — | 34.81 | |
| DRUNetnoise level (σ)=252024.02 | — | — | — | 32.6 | |
| DRUNetnoise level (σ)=502024.02 | — | — | — | 29.61 | |
| DRUNetσ=152024.11 | — | — | — | 34.81 | |
| FFDNetnoise level (σ)=152024.02 | — | — | — | 33.83 | |
| FFDNetnoise level (σ)=252024.02 | — | — | — | 31.4 | |
| FFDNetnoise level (σ)=502024.02 | — | — | — | 28.05 | |
| FFDNetσ=152024.11 | — | — | — | 33.83 | |
| IRCNNnoise level (σ)=152024.02 | — | — | — | 33.78 | |
| IRCNNnoise level (σ)=252024.02 | — | — | — | 31.2 | |
| IRCNNnoise level (σ)=502024.02 | — | — | — | 27.7 | |
| IRCNNσ=152024.11 | — | — | — | 33.78 | |
| MambaIRnoise level (σ)=152024.02 | — | — | — | 35.37 | |
| MambaIRnoise level (σ)=252024.02 | — | — | — | 33.21 | |
| MambaIRnoise level (σ)=502024.02 | — | — | — | 30.3 | |
| MambaIRσ=152024.11 | — | — | — | 35.37 | |
| MambaIRv2σ=152024.11 | — | — | — | 35.42 | |
| Restormernoise level (σ)=152024.02 | — | — | — | 35.13 | |
| Restormernoise level (σ)=252024.02 | — | — | — | 32.96 | |
| Restormernoise level (σ)=502024.02 | — | — | — | 30.02 | |
| Restormerσ=152024.11 | — | — | — | 35.13 | |
| SwinIRnoise level (σ)=152024.02 | — | — | — | 35.13 | |
| SwinIRnoise level (σ)=252024.02 | — | — | — | 32.9 | |
| SwinIRnoise level (σ)=502024.02 | — | — | — | 29.82 | |
| SwinIRσ=152024.11 | — | — | — | 35.13 | |
| Xformerσ=152024.11 | — | — | — | 35.29 |