Super-Resolution on Manga109 (test)
40.42PSNRMambaIRv2
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| MambaIRv2Scale=2x2024.11 | 40.42 | — | — | — | — | — | — | |
| ATDScale=2x2024.11 | 40.37 | — | — | — | — | — | — | |
| SuRGe2024.04 | 34.17 | — | — | — | 0.95 | — | — | |
| HMA2024.04 | 33.19 | — | — | — | 0.93 | — | — | |
| DRCT-L2024.04 | 33.14 | — | — | — | 0.93 | — | — | |
| Hi-IR-L2024.04 | 33.13 | — | — | — | 0.93 | — | — | |
| HAT-L2024.04 | 33.09 | — | — | — | 0.93 | — | — | |
| DRCT2024.04 | 32.96 | — | — | — | 0.93 | — | — | |
| HAT2024.04 | 32.87 | — | — | — | 0.93 | — | — | |
| CPAT+2024.04 | 32.85 | — | — | — | 0.93 | — | — | |
| SwinFIR2024.04 | 32.83 | — | — | — | 0.93 | — | — | |
| MAIR+2024.04 | 32.66 | — | — | — | 0.93 | — | — | |
| MAIR2024.04 | 32.46 | — | — | — | 0.93 | — | — | |
| SwinIR+2024.04 | 32.22 | — | — | — | 0.92 | — | — | |
| DBPN-RES2024.04 | 31.74 | — | — | — | 0.92 | — | — | |
| SAN2024.04 | 31.66 | — | — | — | 0.92 | — | — | |
| CAMixerSRScale=x4, #Params=765K, #MAdds=53.8G2024.02 | 31.18 | — | — | — | 0.9166 | — | — | |
| SwinIR-NGScale=x4, #Params=1201K, #MAdds=63.0G2024.02 | 31.09 | — | — | — | 0.9161 | — | — | |
| SwinIR-lightScale=x4, #Params=930K, #MAdds=61.7G2024.02 | 30.92 | — | — | — | 0.9151 | — | — | |
| ELAN-lightScale=x4, #Params=601K, #MAdds=43.2G2024.02 | 30.92 | — | — | — | 0.915 | — | — | |
| NGswinScale=x4, #Params=1019K, #MAdds=36.4G2024.02 | 30.8 | — | — | — | 0.9128 | — | — | |
| DiVANetScale=x4, #Params=939K, #MAdds=57.0G2024.02 | 30.73 | — | — | — | 0.9119 | — | — | |
| DATSR-recSR paradigm=RefSR, Loss=reconstruction loss2022.07 | 30.49 | — | — | — | 0.912 | — | — | |
| C2-Matching-recSR paradigm=RefSR, Loss=reconstruction loss2022.07 | 30.47 | — | — | — | 0.911 | — | — | |
| IMDNScale=x4, #Params=715K, #MAdds=40.9G2024.02 | 30.45 | — | — | — | 0.9075 | — | — | |
| MASA-recSR paradigm=RefSR, Loss=reconstruction loss2022.07 | 30.24 | — | — | — | 0.909 | — | — | |
| SwinIR + LDLBackbone=SwinIR, Training Dataset=DF2K, Scaling Factor=4x2022.03 | 30.143 | 0.0469 | 0.0315 | 8.68 | 0.888 | — | — | |
| TTSR-recSR paradigm=RefSR, Loss=reconstruction loss2022.07 | 30.09 | — | — | — | 0.907 | — | — | |
| SwinIRSR paradigm=SISR, Loss=reconstruction loss2022.07 | 30.05 | — | — | — | 0.91 | — | — | |
| AESOPUpscaling Factor=x42026.06 | 29.973 | 0.0525 | 0.036 | — | 0.8827 | — | — | |
| MaCo-GANUpscaling Factor=x42026.06 | 29.865 | 0.0511 | 0.0362 | — | 0.8794 | — | — | |
| DATSRSR paradigm=RefSR2022.07 | 29.75 | — | — | — | 0.893 | — | — | |
| C2-MatchingSR paradigm=RefSR2022.07 | 29.73 | — | — | — | 0.893 | — | — | |
| RRDB + LDLBackbone=RRDB, Training Dataset=DF2K, Scaling Factor=4x2022.03 | 29.62 | 0.0544 | 0.0355 | 10.161 | 0.8734 | — | — | |
| LDLUpscaling Factor=x42026.06 | 29.62 | 0.0544 | 0.0355 | — | 0.8734 | — | — | |
| RRDB + LDLBackbone=RRDB, Training Dataset=DIV2K, Scaling Factor=4x2022.03 | 29.407 | 0.0553 | 0.0404 | 9.855 | 0.8746 | — | — | |
| RCANSR paradigm=SISR, Loss=reconstruction loss2022.07 | 29.38 | — | — | — | 0.895 | — | — | |
| SwinIR + LGANBackbone=SwinIR, Training Dataset=DF2K, Scaling Factor=4x2022.03 | 29.345 | 0.0542 | 0.0365 | 9.703 | 0.8796 | — | — | |
| SRNTT-recSR paradigm=RefSR, Loss=reconstruction loss2022.07 | 28.95 | — | — | — | 0.885 | — | — | |
| EDSRSR paradigm=SISR, Loss=reconstruction loss2022.07 | 28.93 | — | — | — | 0.891 | — | — | |
| USRGANBackbone=RRDB, Training Dataset=DF2K, Scaling Factor=4x2022.03 | 28.753 | 0.063 | 0.0471 | 10.658 | 0.8717 | — | — | |
| TTSRSR paradigm=RefSR2022.07 | 28.7 | — | — | — | 0.886 | — | — | |
| SRResNet + LDLBackbone=SRResNet-like, Training Dataset=DIV2K, Scaling Factor=4x2022.03 | 28.664 | 0.0673 | 0.0523 | 12.652 | 0.8702 | — | — | |
| SPSRBackbone=RRDB, Training Dataset=DIV2K, Scaling Factor=4x2022.03 | 28.561 | 0.0672 | 0.0463 | 10.662 | 0.859 | — | — | |
| SPSRUpscaling Factor=x42026.06 | 28.561 | 0.0672 | 0.0463 | — | 0.859 | — | — | |
| ESRGANBackbone=RRDB, Training Dataset=DF2K + OST, Scaling Factor=4x2022.03 | 28.413 | 0.0649 | 0.0471 | 11.552 | 0.8595 | — | — | |
| ESRGANUpscaling Factor=x42026.06 | 28.413 | 0.0649 | 0.0471 | — | 0.8595 | — | — | |
| PadéNet-IDNumber of Parameters=≈ 445K, Scale=x42026.01 | 28.28 | — | — | — | 0.894 | 0.1136 | 0.1717 | |
| PadéNetNumber of Parameters=≈ 445K, Scale=x42026.01 | 28.25 | — | — | — | 0.8935 | 0.1146 | 0.1717 | |
| SFTGANBackbone=SRResNet-like, Training Dataset=ImageNet + OST, Scaling Factor=4x2022.03 | 28.167 | 0.0716 | 0.0646 | 21.464 | 0.8562 | — | — | |
| DCN 1 × 1Number of Parameters=≈ 447K, Scale=x42026.01 | 28.16 | — | — | — | 0.8923 | 0.116 | 0.1719 | |
| DCN 3 × 3Number of Parameters=≈ 509K, Scale=x42026.01 | 28.13 | — | — | — | 0.8917 | 0.1162 | 0.1724 | |
| SRGANBackbone=SRResNet-like, Training Dataset=DIV2K, Scaling Factor=4x2022.03 | 28.11 | 0.0707 | 0.0557 | 11.948 | 0.8632 | — | — | |
| SelfONNNumber of Parameters=≈ 440K, Scale=x42026.01 | 28.08 | — | — | — | 0.8908 | 0.1178 | 0.1717 | |
| SuperONNNumber of Parameters=≈ 440K, Scale=x42026.01 | 28.05 | — | — | — | 0.8902 | 0.1174 | 0.172 | |
| ResNetNumber of Parameters=≈ 440K, Scale=x42026.01 | 28.02 | — | — | — | 0.8904 | 0.1186 | 0.1739 | |
| PAU-NetNumber of Parameters=≈ 440K, Scale=x42026.01 | 27.98 | — | — | — | 0.8894 | 0.1188 | 0.1711 | |
| SRCNN2024.04 | 27.66 | — | — | — | 0.86 | — | — | |
| SRNTTSR paradigm=RefSR2022.07 | 27.54 | — | — | — | 0.862 | — | — | |
| MASASR paradigm=RefSR2022.07 | 27.26 | — | — | — | 0.847 | — | — | |
| SRCNNSR paradigm=SISR, Loss=reconstruction loss2022.07 | 27.12 | — | — | — | 0.85 | — | — | |
| ENetSR paradigm=SISR, Loss=reconstruction loss2022.07 | 25.25 | — | — | — | 0.802 | — | — | |
| SRGANSR paradigm=SISR2022.07 | 25.12 | — | — | — | 0.802 | — | — | |
| RankSRGANSR paradigm=SISR2022.07 | 25.04 | — | — | — | 0.803 | — | — | |
| ESRGANSR paradigm=SISR2022.07 | 23.53 | — | — | — | 0.797 | — | — | |
| CrossNetSR paradigm=RefSR, Loss=reconstruction loss2022.07 | 23.36 | — | — | — | 0.741 | — | — |