Semantic Image Synthesis on ADE20K (test)
6.09FIDStable Diffusion
Evaluation Results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Stable Diffusionsegmentation_conditions=false2023.02 | 6.09 | — | — | — | — | — | — | 0.26 | 6.32 | |
| ControlNetsegmentation_conditions=true2023.02 | 15.27 | — | — | — | — | — | — | 0.26 | 6.31 | |
| ControlNet-litesegmentation_conditions=true2023.02 | 17.92 | — | — | — | — | — | — | 0.26 | 6.3 | |
| PITIsegmentation_conditions=true2023.02 | 19.74 | — | — | — | — | — | — | 0.2 | 5.77 | |
| LDMsegmentation_conditions=true, trained_from_scratch=true2023.02 | 25.35 | — | — | — | — | — | — | 0.18 | 5.15 | |
| VQGANsegmentation_conditions=true, trained_from_scratch=true2023.02 | 26.28 | — | — | — | — | — | — | 0.17 | 5.14 | |
| SCAMImage resolution=256px2022.10 | 27.5 | — | — | — | — | — | 20 | — | — | |
| CC-FPSE2019.10 | 31.7 | — | — | — | 43.7 | 82.9 | — | — | — | |
| SPADE + FFLfocal frequency loss (FFL)=true2020.12 | 33.7 | — | — | — | 42.9 | 82.4 | — | — | — | |
| SPADE2019.10 | 33.9 | — | — | — | 38.5 | 79.9 | — | — | — | |
| SPADEfocal frequency loss (FFL)=false2020.12 | 33.9 | — | — | — | 38.5 | 79.9 | — | — | — | |
| SEAN-CLADEImage resolution=256px2022.10 | 38.7 | — | — | — | — | — | 14 | — | — | |
| SEANImage resolution=256px2022.10 | 47.6 | — | — | — | — | — | 14.6 | — | — | |
| INADEImage resolution=256px2022.10 | 48.6 | — | — | — | — | — | 11.3 | — | — | |
| CLADEImage resolution=256px2022.10 | 53.7 | — | — | — | — | — | 10.4 | — | — | |
| SPADEImage resolution=256px2022.10 | 59.7 | — | — | — | — | — | 10.7 | — | — | |
| CRN2019.10 | 73.3 | — | — | — | 22.4 | 68.8 | — | — | — | |
| CRN2020.12 | 73.3 | — | — | — | 22.4 | 68.8 | — | — | — | |
| pix2pixHD2019.10 | 81.8 | — | — | — | 20.3 | 69.2 | — | — | — | |
| pix2pixHD2020.12 | 81.8 | — | — | — | 20.3 | 69.2 | — | — | — | |
| SPADEevaluation_protocol=User preference study (AMT)2019.03 | — | 76.66 | 83.74 | — | — | — | — | — | — |