Super-resolution on ImageNet 1K (test)
27.45PSNRDDNM
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| DDNMNoise level (sigma_e)=0, Upsampling scale=x42023.12 | 27.45 | 0.245 | — | — | — | |
| DDPGNoise level (sigma_e)=0, Upsampling scale=x42023.12 | 27.41 | 0.255 | — | — | — | |
| DDRMNoise level (sigma_e)=0, Upsampling scale=x42023.12 | 27.38 | 0.27 | — | — | — | |
| IDPGNoise level (sigma_e)=0, Upsampling scale=x42023.12 | 27.2 | 0.326 | — | — | — | |
| DiffPIRNoise level (sigma_e)=0, Upsampling scale=x42023.12 | 26.99 | 0.225 | — | — | — | |
| ProjDiffNFEs=100, Noise Level (sigma)=0.052024.06 | 25.73 | 34.12 | 0.73 | 55.03 | — | |
| DDRMNFEs=100, Noise Level (sigma)=0.052024.06 | 25.66 | 34.88 | 0.72 | 55.71 | — | |
| DDNM+NFEs=100, Noise Level (sigma)=0.052024.06 | 25.61 | 34.45 | 0.72 | 54.1 | — | |
| DPSNoise level (sigma_e)=0, Number of Function Evaluations (NFEs)=1000, Upsampling scale=x42023.12 | 25.56 | 0.236 | — | — | — | |
| DDPGNoise level (sigma_e)=0.05, Upsampling scale=x42023.12 | 25.55 | 0.354 | — | — | — | |
| DDRMNoise level (sigma_e)=0.05, Upsampling scale=x42023.12 | 25.54 | 0.333 | — | — | — | |
| IDPGNoise level (sigma_e)=0.05, Upsampling scale=x42023.12 | 25.51 | 0.411 | — | — | — | |
| DiffPIRNoise level (sigma_e)=0.05, Upsampling scale=x42023.12 | 24.65 | 0.318 | — | — | — | |
| DPSNFEs=1000, Noise Level (sigma)=0.052024.06 | 24.44 | 31.81 | 0.67 | 36.17 | — | |
| pMF tiled gradient-free CBG with renoiseScale=4x, Resolution=256x256, Likelihood noise (σy)=0.01, K=2048, denoising steps=8, tiles=64x642026.02 | 24.41 | 0.298 | 0.668 | 40.74 | 0.0033 | |
| DDRMScale=4x, Resolution=256x256, Likelihood noise (σy)=0.012026.02 | 24.15 | 0.325 | 0.701 | 52.76 | 0.0151 | |
| RED-diffScale=4x, Resolution=256x256, Likelihood noise (σy)=0.012026.02 | 24.06 | 0.354 | 0.685 | 51.83 | 0.0084 | |
| DPSNoise level (sigma_e)=0.05, Number of Function Evaluations (NFEs)=1000, Upsampling scale=x42023.12 | 24.05 | 0.271 | — | — | — | |
| DPSScale=4x, Resolution=256x256, Likelihood noise (σy)=0.012026.02 | 23.61 | 0.195 | 0.676 | 30.67 | 0.0021 | |
| RB-ModulationScale=4x, Resolution=256x256, Likelihood noise (σy)=0.012026.02 | 23.41 | 0.211 | 0.674 | 35.26 | 0.0032 | |
| C-ΠGDMScale=4x, Resolution=256x256, Likelihood noise (σy)=0.012026.02 | 23.2 | 0.27 | 0.631 | 39.96 | 0.0024 | |
| NDTMScale=4x, Resolution=256x256, Likelihood noise (σy)=0.012026.02 | 23.12 | 0.158 | 0.674 | 28.75 | 0.0011 | |
| RED-diffNFEs=100, Noise Level (sigma)=0.052024.06 | 22.74 | 53.24 | 0.49 | 96.26 | — | |
| A†yNoise Level (sigma)=0.052024.06 | 21.85 | 65.34 | 0.45 | 183.32 | — | |
| MPGDScale=4x, Resolution=256x256, Likelihood noise (σy)=0.012026.02 | 21.83 | 0.215 | 0.587 | 37.39 | 0.0017 |