Semantic Image Synthesis on COCO-Stuff (val)
13.6FIDDP-SIMS
Evaluation Results
| Method | Links | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| DP-SIMSBackbone=ConvNeXt-L2023.12 | 13.6 | 57.7 | — | — | — | — | — | — | — | — | 6,520 | |
| PLACEArchitecture=Diffusion-based2024.03 | 14 | 42.6 | — | — | — | — | — | — | — | — | — | |
| FreestyleNetArchitecture=Diffusion-based2024.03 | 14.4 | 40.7 | — | — | — | — | — | — | — | — | — | |
| PITI2023.12 | 15.5 | 29.5 | — | — | — | — | — | — | — | — | 3,120 | |
| ECGANArchitecture=GAN-based2024.03 | 15.7 | 46.3 | — | — | — | — | — | — | — | — | — | |
| PoE-GAN2023.12 | 15.8 | — | — | — | — | — | — | — | — | — | — | |
| SDM2023.12 | 15.9 | 36.8 | — | — | — | — | — | — | — | — | 4,030 | |
| SDMArchitecture=Diffusion-based2024.03 | 15.9 | 40.2 | — | — | — | — | — | — | — | — | — | |
| PITIArchitecture=Diffusion-based2024.03 | 16.1 | 34.1 | — | — | — | — | — | — | — | — | — | |
| T2I-AdapterArchitecture=Diffusion-based2024.03 | 16.8 | 20.7 | — | — | — | — | — | — | — | — | — | |
| OASISNumber of parameters=94M, VGG Perceptual Loss=false2020.12 | 17 | 44.1 | — | — | — | — | — | — | — | — | — | |
| OASIS2023.12 | 17 | 44.1 | — | — | — | — | — | — | — | — | 5,210 | |
| OASISArchitecture=GAN-based2024.03 | 17 | 44.1 | — | — | — | — | — | — | — | — | — | |
| SC-GANArchitecture=GAN-based2024.03 | 18.1 | 42 | — | — | — | — | — | — | — | — | — | |
| RESAILArchitecture=GAN-based2024.03 | 18.3 | 44.7 | — | — | — | — | — | — | — | — | — | |
| CC-FPSEedge_loss=true, warping_augmentation=false2020.11 | 18.6 | 41.8 | 46.4 | 40.9 | 71.1 | 76.7 | 68.7 | 52.8 | 61.2 | 51.1 | — | |
| CC-FPSEAugmentation=None2020.11 | 18.9 | 41 | 47.3 | 39.7 | — | — | — | — | — | — | — | |
| CC-FPSEedge_loss=false, warping_augmentation=false2020.11 | 18.9 | 41 | 47.3 | 39.7 | 71 | 77.3 | 68.4 | 52.4 | 61.5 | 50.5 | — | |
| CC-FPSE-AUGAugmentation=Label map warping + Edge loss2020.11 | 19.1 | 42.1 | 46.3 | 41.2 | — | — | — | — | — | — | — | |
| CC-FPSEedge_loss=true, warping_augmentation=true2020.11 | 19.1 | 42.1 | 46.3 | 41.2 | 71.5 | 78.1 | 68.8 | 53.1 | 61.8 | 51.3 | — | |
| CC-FPSENumber of parameters=131M, VGG Perceptual Loss=true2020.12 | 19.2 | 41.6 | — | — | — | — | — | — | — | — | — | |
| CC-FPSEArchitecture=GAN-based2024.03 | 19.2 | 41.6 | — | — | — | — | — | — | — | — | — | |
| SPADE+Number of parameters=102M, VGG Perceptual Loss=true2020.12 | 21.7 | 38.8 | — | — | — | — | — | — | — | — | — | |
| SPADEAugmentation=None2020.11 | 22.5 | 37.8 | 43.5 | 36.7 | — | — | — | — | — | — | — | |
| SPADEedge_loss=false, warping_augmentation=false2020.11 | 22.5 | 37.8 | 43.5 | 36.7 | 68.6 | 74.5 | 66.2 | 48.8 | 56.9 | 47.1 | — | |
| SPADENumber of parameters=102M, VGG Perceptual Loss=true2020.12 | 22.6 | 37.4 | — | — | — | — | — | — | — | — | — | |
| SPADE2023.12 | 22.6 | 37.4 | — | — | — | — | — | — | — | — | — | |
| SPADEArchitecture=GAN-based2024.03 | 22.6 | 37.4 | — | — | — | — | — | — | — | — | — | |
| SPADE-AUGAugmentation=Label map warping + Edge loss2020.11 | 22.7 | 38.2 | 43.5 | 37.1 | — | — | — | — | — | — | — | |
| SPADEedge_loss=true, warping_augmentation=true2020.11 | 22.7 | 38.2 | 43.5 | 37.1 | 68.8 | 75.3 | 66.1 | 49.4 | 57.4 | 47.7 | — | |
| SPADEedge_loss=true, warping_augmentation=false2020.11 | 22.8 | 37.8 | 43.4 | 36.7 | 68.3 | 73.1 | 66.3 | 48.9 | 57.6 | 47.1 | — | |
| SAFMArchitecture=GAN-based2024.03 | 24.6 | 43.3 | — | — | — | — | — | — | — | — | — | |
| Pix2PixHD-AUGAugmentation=Label map warping + Edge loss2020.11 | 54.2 | 21.9 | 31.8 | 19.9 | — | — | — | — | — | — | — | |
| Pix2PixHDedge_loss=true, warping_augmentation=true2020.11 | 54.2 | 21.9 | 31.8 | 19.9 | 54.1 | 69.2 | 47.8 | 29.5 | 46.8 | 26 | — | |
| CRNNumber of parameters=84M, VGG Perceptual Loss=true2020.12 | 70.4 | 23.7 | — | — | — | — | — | — | — | — | — | |
| Pix2PixHDedge_loss=true, warping_augmentation=false2020.11 | 91.2 | 17.9 | 27.3 | 15.9 | 43.9 | 55.7 | 39 | 25.5 | 37.7 | 23 | — | |
| SPADE+Number of parameters=102M, VGG Perceptual Loss=false2020.12 | 99.1 | 16.1 | — | — | — | — | — | — | — | — | — | |
| Pix2pixHDNumber of parameters=183M, VGG Perceptual Loss=true2020.12 | 111.5 | 14.6 | — | — | — | — | — | — | — | — | — | |
| Pix2pixHD2023.12 | 111.5 | 14.6 | — | — | — | — | — | — | — | — | — | |
| pix2pixHDArchitecture=GAN-based2024.03 | 111.5 | 14.6 | — | — | — | — | — | — | — | — | — | |
| Pix2PixHDAugmentation=None2020.11 | 128.7 | 12 | 21.2 | 10.1 | — | — | — | — | — | — | — | |
| Pix2PixHDedge_loss=false, warping_augmentation=false2020.11 | 128.7 | 12 | 21.2 | 10.1 | 36.9 | 47.6 | 32.5 | 18 | 28.7 | 15.8 | — |