Super-Resolution on RealSR (test)
34.35PSNRMIRNet
Evaluation results
| Method | Links | |||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| MIRNetScale=x22020.03 | 34.35 | 0.935 | — | — | — | — | — | — | — | — | — | — | — | |
| LP-KPNScale=x22020.03 | 33.9 | 0.927 | — | — | — | — | — | — | — | — | — | — | — | |
| RCANScale=x22020.03 | 33.87 | 0.922 | — | — | — | — | — | — | — | — | — | — | — | |
| SRResNetScale=x22020.03 | 33.69 | 0.919 | — | — | — | — | — | — | — | — | — | — | — | |
| VDSRScale=x22020.03 | 33.64 | 0.917 | — | — | — | — | — | — | — | — | — | — | — | |
| BicubicScale=x22020.03 | 32.61 | 0.907 | — | — | — | — | — | — | — | — | — | — | — | |
| MIRNetScale=x32020.03 | 31.16 | 0.885 | — | — | — | — | — | — | — | — | — | — | — | |
| MIRNet-v2Scale=x32022.04 | 31.15 | 0.883 | — | — | — | — | — | — | — | — | — | — | — | |
| LP-KPNScale=x32020.03 | 30.42 | 0.868 | — | — | — | — | — | — | — | — | — | — | — | |
| LP-KPNScale=x32022.04 | 30.42 | 0.868 | — | — | — | — | — | — | — | — | — | — | — | |
| RCANScale=x32020.03 | 30.4 | 0.862 | — | — | — | — | — | — | — | — | — | — | — | |
| RCANScale=x32022.04 | 30.4 | 0.862 | — | — | — | — | — | — | — | — | — | — | — | |
| SRResNetScale=x32020.03 | 30.18 | 0.859 | — | — | — | — | — | — | — | — | — | — | — | |
| SRResNetScale=x32022.04 | 30.18 | 0.859 | — | — | — | — | — | — | — | — | — | — | — | |
| VDSRScale=x32020.03 | 30.14 | 0.856 | — | — | — | — | — | — | — | — | — | — | — | |
| VDSRScale=x32022.04 | 30.14 | 0.856 | — | — | — | — | — | — | — | — | — | — | — | |
| BicubicScale=x32020.03 | 29.34 | 0.841 | — | — | — | — | — | — | — | — | — | — | — | |
| BicubicScale=x32022.04 | 29.34 | 0.841 | — | — | — | — | — | — | — | — | — | — | — | |
| CFSRParams.=307K, FLOPS=17.5G2024.01 | 29.25 | 0.8266 | — | — | — | — | — | — | — | — | — | — | — | |
| ShuffleMixerParams.=411K, FLOPS=28.0G2024.01 | 29.16 | 0.8261 | — | — | — | — | — | — | — | — | — | — | — | |
| MIRNetScale=x42020.03 | 29.14 | 0.843 | — | — | — | — | — | — | — | — | — | — | — | |
| LP-KPNScale=x42020.03 | 28.92 | 0.834 | — | — | — | — | — | — | — | — | — | — | — | |
| RCANScale=x42020.03 | 28.88 | 0.826 | — | — | — | — | — | — | — | — | — | — | — | |
| SRResNetScale=x42020.03 | 28.67 | 0.824 | — | — | — | — | — | — | — | — | — | — | — | |
| VDSRScale=x42020.03 | 28.63 | 0.821 | — | — | — | — | — | — | — | — | — | — | — | |
| BicubicScale=x42020.03 | 27.99 | 0.806 | — | — | — | — | — | — | — | — | — | — | — | |
| PASDScale=x42026.04 | 26.792 | 0.762 | 0.215 | — | 4.867 | 58.689 | — | — | — | 0.521 | 5.546 | 4.77 | — | |
| PASD2026.03 | 26.63 | 0.766 | 0.2871 | — | — | 59.99 | 0.3939 | 0.4855 | — | — | — | — | — | |
| BSRGAN2026.06 | 26.39 | 0.7654 | 0.267 | 0.2121 | 5.66 | 63.21 | 0.5399 | — | 141.28 | 0.5001 | — | — | — | |
| BSRGAN2026.03 | 26.38 | 0.7651 | 0.2656 | — | — | 63.28 | 0.3758 | 0.5114 | — | — | — | — | — | |
| SinSR2026.05 | 26.32 | 0.7363 | 0.3195 | 0.2351 | 6.3541 | 60.42 | 0.5366 | — | 138.64 | — | — | — | 0.6153 | |
| ResShift#Steps=152026.05 | 26.31 | 0.7421 | 0.346 | 0.2498 | 7.26 | 58.43 | 0.5285 | — | 141.71 | 0.5444 | — | — | — | |
| ResShiftVenue=NeurIPS’23, Step strategy=Prior-based multi-step / flexible-step2026.05 | 26.31 | 0.7421 | 0.346 | 0.2498 | 7.2635 | 58.43 | 0.5285 | — | — | 0.5444 | — | — | — | |
| SinSR2026.03 | 26.3 | 0.7354 | 0.3212 | — | — | 60.41 | 0.5389 | — | 137.05 | 0.6204 | — | — | — | |
| SinSRSampling Steps=12026.06 | 26.3 | 0.7354 | 0.3212 | 0.2346 | 6.31 | 60.41 | 0.5389 | — | 137.05 | 0.6204 | — | — | — | |
| SinSRSampling strategy=One-step2025.12 | 26.28 | 0.7351 | 0.3189 | 0.2347 | 6.31 | 60.64 | 0.5384 | 0.4401 | — | — | — | — | — | |
| SinSR#Steps=12026.05 | 26.28 | 0.7347 | 0.3188 | 0.2353 | 6.29 | 60.8 | 0.5385 | — | 135.93 | 0.6122 | — | — | — | |
| FreqOrtho-SR2026.06 | 26.27 | 0.7481 | 0.2537 | 0.1951 | 5.32 | 69.39 | 0.6586 | — | 108.91 | 0.663 | — | — | — | |
| FreqOrtho-SRSampling Steps=12026.06 | 26.27 | 0.7481 | 0.2537 | 0.1951 | 5.32 | 69.39 | 0.6586 | — | 108.91 | 0.663 | — | — | — | |
| FreqOrtho-SRSteps=Single-step2026.06 | 26.27 | 0.7481 | 0.2537 | 0.1951 | 5.32 | 69.39 | 0.6586 | — | 108.91 | 0.663 | — | — | — | |
| MagicBokehDefocus Map=all-zero input2026.05 | 26.14 | 0.7392 | 0.2888 | 0.2192 | 5.6337 | 67.22 | 0.6214 | — | 123.06 | — | — | — | 0.6246 | |
| SinSRVenue=CVPR’24, Step strategy=Efficient / one-step2026.05 | 25.98 | 0.7347 | 0.3188 | 0.2353 | 6.2872 | 60.8 | 0.5385 | — | — | 0.6122 | — | — | — | |
| SANA-SRStep strategy=Efficient / one-step2026.05 | 25.96 | 0.7468 | 0.2917 | 0.1964 | 5.0897 | 70.73 | 0.5932 | — | — | 0.7663 | — | — | — | |
| Real-ESRGANScale=x42026.04 | 25.845 | 0.774 | 0.215 | — | 4.679 | 59.696 | — | — | — | 0.49 | 5.868 | 4.489 | — | |
| TVTSampling strategy=One-step2025.12 | 25.81 | 0.7596 | 0.2597 | 0.2061 | 5.92 | 69.89 | 0.6232 | 0.548 | — | — | — | — | — | |
| TVTSampling Steps=12026.06 | 25.81 | 0.7596 | 0.2587 | 0.2061 | 5.92 | 69.89 | 0.6228 | — | 109.93 | 0.6882 | — | — | — | |
| SeeSRScale=x42026.04 | 25.799 | 0.741 | 0.223 | — | 4.005 | 66.983 | — | — | — | 0.675 | 6.469 | 3.852 | — | |
| RealESRGAN2026.06 | 25.69 | 0.7616 | 0.2727 | 0.2063 | 5.83 | 60.18 | 0.5487 | — | 135.18 | 0.4449 | — | — | — | |
| DoSSR#Steps=12026.05 | 25.67 | 0.7387 | 0.3356 | 0.2741 | 11.19 | 62.69 | 0.5243 | — | 153.25 | 0.6259 | — | — | — | |
| PiSA-SRScale=x42026.04 | 25.587 | 0.749 | 0.208 | — | 4.388 | 67.995 | — | — | — | 0.668 | 6.421 | 4.033 | — | |
| Real-ESRGAN2026.03 | 25.56 | 0.7571 | 0.2688 | — | — | 60.15 | 0.3788 | 0.4559 | — | — | — | — | — | |
| QUSR2026.03 | 25.54 | 0.7289 | 0.2974 | — | — | 69.17 | 0.6564 | — | 125.27 | 0.6824 | — | — | — | |
| FlowSR#Steps=12026.05 | 25.54 | 0.7434 | 0.2728 | 0.2013 | 5.28 | 69.22 | 0.6486 | — | 112.6 | 0.6701 | — | — | — | |
| PiSA-SRSampling strategy=One-step2025.12 | 25.5 | 0.7418 | 0.2672 | 0.2044 | 5.51 | 70.15 | 0.6552 | 0.5635 | — | — | — | — | — | |
| PiSA-SR2026.03 | 25.5 | 0.7414 | 0.2672 | — | — | 70.15 | 0.656 | — | 124.09 | 0.6702 | — | — | — | |
| PiSA-SRSampling Steps=12026.06 | 25.5 | 0.7417 | 0.2672 | 0.2044 | 5.5 | 70.15 | 0.656 | — | 124.09 | 0.6702 | — | — | — | |
| ResShiftScale=x42026.04 | 25.453 | 0.725 | 0.266 | — | 7.349 | 56.189 | — | — | — | 0.585 | 6.202 | 5.55 | — | |
| CODSRSampling strategy=One-step2025.12 | 25.37 | 0.7284 | 0.2741 | 0.2002 | 5.31 | 70.54 | 0.6727 | 0.5768 | — | — | — | — | — | |
| SeeSRSampling strategy=Full-step2025.12 | 25.34 | 0.7294 | 0.2961 | 0.2197 | 5.35 | 69.45 | 0.6453 | 0.5762 | — | — | — | — | — | |
| SeeSR2026.03 | 25.33 | 0.7273 | 0.2985 | — | — | 69.37 | 0.6439 | — | 125.66 | 0.6594 | — | — | — | |
| SeeSR-S50sampling steps=502026.06 | 25.33 | 0.7273 | 0.2985 | 0.2213 | 5.38 | 69.37 | 0.6439 | — | 125.66 | 0.6594 | — | — | — | |
| MoR-DASRSampling Steps=12026.06 | 25.32 | 0.728 | 0.291 | — | — | 69.78 | 0.512 | — | — | 0.691 | — | — | — | |
| AdcSRVenue=CVPR’25, Step strategy=Efficient / one-step2026.05 | 25.31 | 0.7238 | 0.2997 | 0.2162 | 5.3152 | 70.31 | 0.6372 | — | — | 0.7363 | — | — | — | |
| LDL2026.06 | 25.28 | 0.7567 | 0.2766 | 0.2121 | 6 | 60.82 | 0.5485 | — | 142.71 | 0.4477 | — | — | — | |
| FaithDiffSampling strategy=Full-step2025.12 | 25.27 | 0.7068 | 0.2832 | 0.2076 | 5.46 | 68.96 | 0.6793 | 0.5847 | — | — | — | — | — | |
| OSEDiffScale=x42026.04 | 25.254 | 0.736 | 0.23 | — | 4.342 | 67.597 | — | — | — | 0.677 | 6.532 | 3.969 | — | |
| FaithDiff2026.03 | 25.23 | 0.7069 | 0.2872 | — | — | 68.86 | 0.4644 | 0.6126 | — | — | — | — | — | |
| PASD-S20sampling steps=202026.06 | 25.22 | 0.6809 | 0.3392 | 0.2259 | 5.18 | 68.74 | 0.6461 | — | 123.08 | 0.6502 | — | — | — | |
| PASD#Steps=202026.05 | 25.21 | 0.6798 | 0.338 | 0.226 | 5.41 | 68.75 | 0.6487 | — | 124.29 | 0.662 | — | — | — | |
| PASDVenue=ECCV’24, Step strategy=Prior-based multi-step / flexible-step2026.05 | 25.21 | 0.6798 | 0.338 | 0.226 | 5.4137 | 68.75 | 0.6487 | — | — | 0.662 | — | — | — | |
| S3Diff2026.05 | 25.18 | 0.7269 | 0.2722 | 0.2005 | 5.2612 | 67.82 | 0.6417 | — | 105.11 | — | — | — | 0.6742 | |
| SeeSR#Steps=502026.05 | 25.18 | 0.7216 | 0.3009 | 0.2223 | 5.41 | 69.77 | 0.6442 | — | 125.55 | 0.6612 | — | — | — | |
| SeeSRVenue=CVPR’24, Step strategy=Prior-based multi-step / flexible-step2026.05 | 25.18 | 0.7216 | 0.3009 | 0.2223 | 5.4081 | 69.77 | 0.6442 | — | — | 0.6612 | — | — | — | |
| SUPIRSampling strategy=Full-step2025.12 | 25.16 | 0.6872 | 0.3653 | 0.2439 | 6.96 | 59.3 | 0.5629 | 0.5063 | — | — | — | — | — | |
| OSEDiffSampling strategy=One-step2025.12 | 25.15 | 0.7341 | 0.292 | 0.2127 | 5.64 | 69.08 | 0.6331 | 0.5533 | — | — | — | — | — | |
| SeeSR2026.03 | 25.15 | 0.721 | 0.3007 | — | — | 69.82 | 0.5437 | 0.6701 | — | — | — | — | — | |
| OSEDiff2026.03 | 25.15 | 0.7341 | 0.2921 | — | — | 69.09 | 0.6339 | — | 123.5 | 0.6693 | — | — | — | |
| OSEDiff2026.05 | 25.15 | 0.7341 | 0.2921 | 0.2128 | 5.6528 | 69.11 | 0.6332 | — | 123.68 | — | — | — | 0.6685 | |
| OSEDiff#Steps=12026.05 | 25.15 | 0.7341 | 0.2921 | 0.2128 | 5.65 | 69.09 | 0.6326 | — | 123.49 | 0.6693 | — | — | — | |
| OSEDiffVenue=NeurIPS’24, Step strategy=Efficient / one-step2026.05 | 25.15 | 0.7341 | 0.2921 | 0.2128 | 5.6476 | 69.09 | 0.6326 | — | — | 0.6693 | — | — | — | |
| OSEDiffSampling Steps=12026.06 | 25.15 | 0.7341 | 0.2921 | 0.2128 | 5.65 | 69.09 | 0.6339 | — | 123.5 | 0.6693 | — | — | — | |
| StableSRScale=x42026.04 | 25.071 | 0.78 | 0.198 | — | 6.158 | 52.429 | — | — | — | 0.453 | 4.64 | 5.809 | — | |
| SUPIR2026.03 | 25.04 | 0.6709 | 0.3715 | — | — | 58.51 | 0.4429 | 0.6357 | — | — | — | — | — | |
| S3DiffVenue=ArXiv’24, Step strategy=Efficient / one-step2026.05 | 25.03 | 0.7321 | 0.2699 | 0.1996 | 5.3311 | 67.89 | 0.6263 | — | — | 0.6722 | — | — | — | |
| DiffBIRSampling strategy=Full-step2025.12 | 24.97 | 0.6696 | 0.3486 | 0.2374 | 5.71 | 68.64 | 0.6436 | 0.5757 | — | — | — | — | — | |
| InvSRScale=x42026.04 | 24.928 | 0.734 | 0.211 | — | 4.75 | 66.371 | — | — | — | 0.671 | 6.672 | 4.06 | — | |
| DiffBIR-S50sampling steps=502026.06 | 24.88 | 0.6673 | 0.3567 | 0.229 | 5.63 | 64.66 | 0.6231 | — | 124.56 | 0.6412 | — | — | — | |
| FLUX-SRVenue=ICML’25, Step strategy=Efficient / one-step2026.05 | 24.83 | 0.7375 | 0.3141 | 0.2264 | 5.2097 | 70.08 | 0.6511 | — | — | 0.7381 | — | — | — | |
| TSD-SRVenue=CVPR’25, Step strategy=Efficient / one-step2026.05 | 24.81 | 0.7172 | 0.2743 | 0.2104 | 5.1298 | 70.49 | 0.6347 | — | — | 0.716 | — | — | — | |
| DiffBIR#Steps=502026.05 | 24.75 | 0.6567 | 0.3636 | 0.2312 | 5.53 | 64.98 | 0.6246 | — | 128.99 | 0.6463 | — | — | — | |
| DiffBIRVenue=ECCV’24, Step strategy=Prior-based multi-step / flexible-step2026.05 | 24.75 | 0.6567 | 0.3636 | 0.2312 | 5.5346 | 64.98 | 0.6246 | — | — | 0.6463 | — | — | — | |
| StableSR#Steps=2002026.05 | 24.7 | 0.7085 | 0.3018 | 0.2288 | 5.91 | 65.78 | 0.6221 | — | 128.51 | 0.6178 | — | — | — | |
| StableSRVenue=IJCV’24, Step strategy=Prior-based multi-step / flexible-step2026.05 | 24.7 | 0.7085 | 0.3018 | 0.2288 | 5.9122 | 65.78 | 0.6221 | — | — | 0.6178 | — | — | — | |
| StableSR2026.03 | 24.69 | 0.7052 | 0.3091 | — | — | 65.42 | 0.6211 | — | 127.2 | 0.6195 | — | — | — | |
| DiffBIR2026.03 | 24.65 | 0.6392 | 0.3667 | — | — | 68.99 | 0.5589 | 0.703 | — | — | — | — | — | |
| D3SRVenue=NeurIPS’25, Step strategy=Efficient / one-step2026.05 | 24.54 | 0.727 | 0.305 | 0.2106 | 5.096 | 68.69 | 0.6382 | — | — | 0.6711 | — | — | — | |
| InvSRVenue=CVPR’25, Step strategy=Prior-based multi-step / flexible-step2026.05 | 24.5 | 0.726 | 0.2978 | 0.2492 | 5.2189 | 69.67 | 0.4456 | — | — | 0.6918 | — | — | — | |
| StableSR2026.03 | 23.84 | 0.705 | 0.2969 | — | — | 58.12 | 0.3451 | 0.5707 | — | — | — | — | — | |
| LinearSRVenue=ICLR’26, Step strategy=Prior-based multi-step / flexible-step2026.05 | 23.84 | 0.6848 | 0.3128 | 0.2935 | 5.8509 | 69.39 | 0.6108 | — | — | 0.6731 | — | — | — | |
| SUPIRVenue=CVPR’24, Step strategy=Prior-based multi-step / flexible-step2026.05 | 23.65 | 0.662 | 0.3541 | 0.2488 | 6.1099 | 62.09 | 0.578 | — | — | 0.6707 | — | — | — |