Image Super-resolution on BSD100 (test)
33.12PSNRWaveMixSR-V2
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| WaveMixSR-V2Training dataset=DIV2K, Scale=2x2024.09 | 33.12 | 93.26 | — | — | — | |
| WaveMixSRTraining dataset=DIV2K, Scale=2x2024.09 | 33.08 | 93.22 | — | — | — | |
| DRCT-LScale=x2, Training Dataset=DF2K, Progressive training strategy=true2024.03 | 32.76 | 90.68 | — | — | — | |
| HAT-LTraining dataset=DF2K, Pre-training=ImageNet, Scale=2x2024.09 | 32.74 | 90.66 | — | — | — | |
| SwinFIRTraining dataset=DF2K, Pre-training=ImageNet, Scale=2x2024.09 | 32.64 | 90.54 | — | — | — | |
| HATTraining dataset=DF2K, Scale=2x2024.09 | 32.62 | 90.53 | — | — | — | |
| DRCTScale=x2, Training Dataset=DF2K2024.03 | 32.62 | 90.51 | — | — | — | |
| SwinIRTraining dataset=DF2K, Scale=2x2024.09 | 32.53 | 90.41 | — | — | — | |
| EDTTraining dataset=DF2K, Scale=2x2024.09 | 32.52 | 90.41 | — | — | — | |
| RCN-itTraining dataset=DF2K, Scale=2x2024.09 | 32.48 | 90.34 | — | — | — | |
| NLSNTraining dataset=DIV2K, Scale=2x2024.09 | 32.43 | 90.27 | — | — | — | |
| SANTraining dataset=DIV2K, Scale=2x2024.09 | 32.42 | 90.28 | — | — | — | |
| RCANTraining dataset=DIV2K, Scale=2x2024.09 | 32.41 | 90.27 | — | — | — | |
| IGNNTraining dataset=DIV2K, Scale=2x2024.09 | 32.41 | 90.25 | — | — | — | |
| HANTraining dataset=DIV2K, Scale=2x2024.09 | 32.41 | 90.27 | — | — | — | |
| SwinOIR+Scale=x2, #Params (K)=8412023.03 | 32.34 | 90.22 | — | — | — | |
| DUKDScale=x2, Backbone=RCAN2023.09 | 32.33 | 90.16 | — | — | — | |
| CRAFTScale=x2, #Params=737K2023.08 | 32.33 | 90.16 | — | — | — | |
| EDSRTraining dataset=DIV2K, Scale=2x2024.09 | 32.32 | 90.13 | — | — | — | |
| SwinIRScale=x2, #Params (K)=8782023.03 | 32.31 | 90.12 | — | — | — | |
| SwinIR-lightScale=x2, #Params=878K2023.08 | 32.31 | 90.12 | — | — | — | |
| ELAN-lightScale=x2, #Params=582K2023.08 | 32.3 | 90.12 | — | — | — | |
| KDScale=x2, Backbone=RCAN2023.09 | 32.29 | 90.1 | — | — | — | |
| RKDScale=x2, Backbone=RCAN2023.09 | 32.29 | 90.11 | — | — | — | |
| FAKD*Scale=x2, Backbone=RCAN2023.09 | 32.29 | 90.11 | — | — | — | |
| SwinOIRScale=x2, #Params (K)=8412023.03 | 32.27 | 90.14 | — | — | — | |
| ScratchScale=x2, Backbone=RCAN2023.09 | 32.26 | 90.07 | — | — | — | |
| LatticeNetScale=x2, #Params (K)=7562023.03 | 32.25 | 90.05 | — | — | — | |
| ATScale=x2, Backbone=RCAN2023.09 | 32.25 | 90.05 | — | — | — | |
| HPUN-LScale=x2, #Params=714K2023.08 | 32.25 | 90.06 | — | — | — | |
| ESRTScale=x2, #Params=777K2023.08 | 32.25 | 90.01 | — | — | — | |
| LatticeNetScale=x2, #Params=756K2023.08 | 32.2 | 89.99 | — | — | — | |
| IMDNScale=x2, #Params (K)=6942023.03 | 32.19 | 89.96 | — | — | — | |
| LAPAR-AScale=x2, #Params (K)=5482023.03 | 32.19 | 89.99 | — | — | — | |
| FitNetScale=x2, Backbone=RCAN2023.09 | 32.19 | 89.99 | — | — | — | |
| IMDNScale=x2, #Params=694K2023.08 | 32.19 | 89.96 | — | — | — | |
| LAPAR-AScale=x2, #Params=548k2023.08 | 32.19 | 89.99 | — | — | — | |
| EDSR-baselineScale=x2, #Params=1370K2023.08 | 32.16 | 89.94 | — | — | — | |
| FALSR-AScale=x2, #Params (K)=1,0212023.03 | 32.1 | 89.87 | — | — | — | |
| CARNScale=x2, #Params (K)=1,5922023.03 | 32.09 | 89.78 | — | — | — | |
| CARNScale=x2, #Params=1592K2023.08 | 32.09 | 89.78 | — | — | — | |
| SRMDNFScale Factor=x2, Degradation=bicubic2017.12 | 32.05 | 89.8 | — | — | — | |
| DRRNScale Factor=x2, Degradation=bicubic2017.12 | 32.01 | 89.7 | — | — | — | |
| RED30Scale=x22016.06 | 31.99 | 89.74 | — | — | — | |
| VDSRScale=x22016.06 | 31.9 | 89.6 | — | — | — | |
| SRMDScale Factor=x2, Degradation=bicubic2017.12 | 31.9 | 89.6 | — | — | — | |
| VDSRScale Factor=x2, Degradation=bicubic2017.12 | 31.89 | 89.6 | — | — | — | |
| DRCNScale=x22016.06 | 31.85 | 89.42 | — | — | — | |
| LapSRNScale Factor=x2, Degradation=bicubic2017.12 | 31.8 | 89.5 | — | — | — | |
| SEDSR-GradResiduallearning rate=0.0003, averaging protocol=final 25% evaluations over 3 random seeds, alpha=32026.02 | 31.73 | — | — | — | — | |
| EDSR-GradResiduallearning rate=0.0003, averaging protocol=final 25% evaluations over 3 random seeds, alpha=-32026.02 | 31.7 | — | — | — | — | |
| EDSRlearning rate=0.0003, averaging protocol=final 25% evaluations over 3 random seeds2026.02 | 31.68 | — | — | — | — | |
| SEDSRlearning rate=0.0003, averaging protocol=final 25% evaluations over 3 random seeds2026.02 | 31.63 | — | — | — | — | |
| SRCNNScale Factor=x2, Degradation=bicubic2017.12 | 31.34 | 88.7 | — | — | — | |
| GT + USRNetScale=x22021.03 | 31.34 | 87.72 | — | — | — | |
| KANONoise factor (N)=0, Sampling scale=22025.12 | 30.35 | — | — | — | — | |
| EDSRNoise factor (N)=0, Sampling scale=22025.12 | 29.8 | — | — | — | — | |
| DRCTScale=x3, Training Dataset=DF2K2024.03 | 29.68 | 81.82 | — | — | — | |
| DRCT-LScale=x3, Training Dataset=DF2K, Progressive training strategy=true2024.03 | 29.68 | 82.11 | — | — | — | |
| BicubicScale Factor=x2, Degradation=bicubic2017.12 | 29.55 | 84.3 | — | — | — | |
| SwinOIR+Scale=x3, #Params (K)=1,0252023.03 | 29.27 | 81.11 | — | — | — | |
| DUKDScale=x3, Backbone=RCAN2023.09 | 29.25 | 80.88 | — | — | — | |
| CRAFTScale=x3, #Params=744K2023.08 | 29.24 | 80.93 | — | — | — | |
| SwinOIRScale=x3, #Params (K)=1,0252023.03 | 29.21 | 80.9 | — | — | — | |
| KDScale=x3, Backbone=RCAN2023.09 | 29.21 | 80.8 | — | — | — | |
| RKDScale=x3, Backbone=RCAN2023.09 | 29.21 | 80.8 | — | — | — | |
| FAKD*Scale=x3, Backbone=RCAN2023.09 | 29.21 | 80.79 | — | — | — | |
| ELAN-lightScale=x3, #Params=590K2023.08 | 29.21 | 80.81 | — | — | — | |
| SwinIRScale=x3, #Params (K)=8862023.03 | 29.2 | 80.82 | — | — | — | |
| SwinIR-lightScale=x3, #Params=886K2023.08 | 29.2 | 80.82 | — | — | — | |
| ScratchScale=x3, Backbone=RCAN2023.09 | 29.18 | 80.74 | — | — | — | |
| HPUN-LScale=x3, #Params=723K2023.08 | 29.18 | 80.72 | — | — | — | |
| ATScale=x3, Backbone=RCAN2023.09 | 29.17 | 80.7 | — | — | — | |
| LatticeNetScale=x3, #Params (K)=7652023.03 | 29.15 | 80.59 | — | — | — | |
| LatticeNetScale=x3, #Params=765K2023.08 | 29.15 | 80.59 | — | — | — | |
| ESRTScale=x3, #Params=770K2023.08 | 29.15 | 80.63 | — | — | — | |
| LBNetScale=x3, #Params=736K2023.08 | 29.13 | 80.61 | — | — | — | |
| LAPAR-AScale=x3, #Params (K)=5442023.03 | 29.11 | 80.54 | — | — | — | |
| LAPAR-AScale=x3, #Params=544k2023.08 | 29.11 | 80.54 | — | — | — | |
| IMDNScale=x3, #Params (K)=7032023.03 | 29.09 | 80.46 | — | — | — | |
| EDSR-baselineScale=x3, #Params=1555K2023.08 | 29.09 | 80.52 | — | — | — | |
| IMDNScale=x3, #Params=703K2023.08 | 29.09 | 80.46 | — | — | — | |
| CARNScale=x3, #Params (K)=1,5922023.03 | 29.06 | 80.34 | — | — | — | |
| CARNScale=x3, #Params=1592K2023.08 | 29.06 | 80.34 | — | — | — | |
| FitNetScale=x3, Backbone=RCAN2023.09 | 29.05 | 80.44 | — | — | — | |
| SRMDNFScale Factor=x3, Degradation=bicubic2017.12 | 28.97 | 80.3 | — | — | — | |
| RED30Scale=x32016.06 | 28.93 | 79.94 | — | — | — | |
| DRRNScale Factor=x3, Degradation=bicubic2017.12 | 28.91 | 79.9 | — | — | — | |
| SRMDScale Factor=x3, Degradation=bicubic2017.12 | 28.87 | 79.9 | — | — | — | |
| VDSRScale=x32016.06 | 28.82 | 79.76 | — | — | — | |
| VDSRScale Factor=x3, Degradation=bicubic2017.12 | 28.82 | 79.8 | — | — | — | |
| LapSRNScale Factor=x3, Degradation=bicubic2017.12 | 28.82 | 79.8 | — | — | — | |
| GT + USRNetScale=x32021.03 | 28.82 | 79.35 | — | — | — | |
| DRCNScale=x32016.06 | 28.8 | 79.63 | — | — | — | |
| DIP-FKP + USRNetScale=x22021.03 | 28.61 | 82.06 | — | — | — | |
| SRCNNScale Factor=x3, Degradation=bicubic2017.12 | 28.4 | 78.6 | — | — | — | |
| KANONoise factor (N)=5, Sampling scale=22025.12 | 28.38 | — | — | — | — | |
| DRCTScale=x4, Training Dataset=DF2K2024.03 | 28.18 | 75.32 | — | — | — | |
| DRCT-LScale=x4, Training Dataset=DF2K, Progressive training strategy=true2024.03 | 28.16 | 75.77 | — | — | — | |
| KANONoise factor (N)=0, Sampling scale=32025.12 | 28.1 | — | — | — | — |