Image Reconstruction on CelebA (test)
0.997SSIMSTAF
Evaluation results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| STAFτ=5, training iterations=2502025.02 | 0.997 | — | — | 70.932 | — | — | 0.0005 | — | — | |
| INCODEtraining iterations=2502025.02 | 0.9964 | — | — | 59.853 | — | — | 0.0001 | — | — | |
| SIRENtraining iterations=2502025.02 | 0.9955 | — | — | 58.251 | — | — | 0.0006 | — | — | |
| GAUSStraining iterations=2502025.02 | 0.9934 | — | — | 47.547 | — | — | 0.0003 | — | — | |
| FINERtraining iterations=2502025.02 | 0.9932 | — | — | 48.96 | — | — | 0.0003 | — | — | |
| MFNtraining iterations=2502025.02 | 0.989 | — | — | 44.124 | — | — | 0.0021 | — | — | |
| WIREtraining iterations=2502025.02 | 0.9773 | — | — | 39.283 | — | — | 0.0082 | — | — | |
| Cold DiffusionImage State=Reconstruction2022.08 | 0.973 | 17.5 | 0.042 | — | — | — | — | — | — | |
| Cold DiffusionImage State=Degraded Image2022.08 | 0.942 | 41.2 | 0.089 | — | — | — | — | — | — | |
| EigenGSTraining Dataset=CelebA, ITER=02025.03 | 0.93 | — | — | 29.9 | — | — | — | — | — | |
| EigenGSTraining Dataset=ImageNet, ITER=02025.03 | 0.91 | — | — | 28.7 | — | — | — | — | — | |
| ReLU + PEtraining iterations=2502025.02 | 0.8899 | — | — | 31.088 | — | — | 0.0865 | — | — | |
| TriplaneNet2026.03 | 0.63 | 63.63 | — | — | — | — | 0.51 | 0.29 | 0.34 | |
| NVB-Face2026.03 | 0.63 | 22.71 | — | — | — | — | 0.49 | 0.22 | 0.46 | |
| PanoHead-PTI2026.03 | 0.62 | 65.88 | — | — | — | — | 0.5 | 0.28 | 0.22 | |
| DiffPortrait3D2026.03 | 0.62 | 67.06 | — | — | — | — | 0.49 | 0.27 | 0.21 | |
| GOAE2026.03 | 0.58 | 71.12 | — | — | — | — | 0.53 | 0.27 | 0.21 | |
| DEAR2024.07 | — | 70.7 | — | — | 0.0526 | — | — | — | — | |
| EigenGSTraining Dataset=CelebA, ITER=1002025.03 | — | — | — | — | — | 28 | — | — | — | |
| EigenGSTraining Dataset=CelebA, ITER=5002025.03 | — | — | — | — | — | 66 | — | — | — | |
| EigenGSTraining Dataset=CelebA, ITER=10002025.03 | — | — | — | — | — | 89 | — | — | — | |
| EigenGSTraining Dataset=CelebA, ITER=50002025.03 | — | — | — | — | — | 98 | — | — | — | |
| EigenGSTraining Dataset=CelebA, ITER=100002025.03 | — | — | — | — | — | 99 | — | — | — | |
| EigenGSTraining Dataset=ImageNet, ITER=1002025.03 | — | — | — | — | — | 10 | — | — | — | |
| EigenGSTraining Dataset=ImageNet, ITER=5002025.03 | — | — | — | — | — | 47 | — | — | — | |
| EigenGSTraining Dataset=ImageNet, ITER=10002025.03 | — | — | — | — | — | 80 | — | — | — | |
| EigenGSTraining Dataset=ImageNet, ITER=50002025.03 | — | — | — | — | — | 98 | — | — | — | |
| EigenGSTraining Dataset=ImageNet, ITER=100002025.03 | — | — | — | — | — | 98 | — | — | — | |
| FactorVAE2024.07 | — | 134.5 | — | — | 0.092 | — | — | — | — | |
| GaussianImageITER=1002025.03 | — | — | — | — | — | 0 | — | — | — | |
| GaussianImageITER=5002025.03 | — | — | — | — | — | 0 | — | — | — | |
| GaussianImageITER=10002025.03 | — | — | — | — | — | 0 | — | — | — | |
| GaussianImageITER=50002025.03 | — | — | — | — | — | 91 | — | — | — | |
| GaussianImageITER=100002025.03 | — | — | — | — | — | 97 | — | — | — | |
| GEM2024.07 | — | 46 | — | — | 0.0483 | — | — | — | — | |
| Learned Initresolution=178x1782022.11 | — | — | — | 30.37 | — | — | — | — | — | |
| Ours (Instance Pattern Composers)resolution=178x1782022.11 | — | — | — | 35.93 | — | — | — | — | — | |
| TransINRresolution=178x1782022.11 | — | — | — | 33.33 | — | — | — | — | — | |
| VAE2024.07 | — | 53.3 | — | — | 0.0514 | — | — | — | — | |
| β-TCVAE2024.07 | — | 139.1 | — | — | 0.1132 | — | — | — | — | |
| β-VAE2024.07 | — | 136.2 | — | — | 0.107 | — | — | — | — |