Synthetic Image Generation on NANO-TEM (synthetic)
0.015Noise stdMS-SSIM VAE
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| MS-SSIM VAEStat=med.2026.06 | 0.015 | 29.17 | 1.197 | |
| MS-SSIM VAEStat=mean2026.06 | 0.019 | 28.81 | 1.196 | |
| DCGANStat=med.2026.06 | 0.026 | 23.77 | 1.131 | |
| patch-based progressive training methodology for DDPMsStat=mean2026.06 | 0.028 | 25.5 | 1.14 | |
| patch-based progressive training methodology for DDPMsStat=med.2026.06 | 0.028 | 25.24 | 1.143 | |
| DCGANStat=mean2026.06 | 0.035 | 23.11 | 1.128 | |
| Original (Target Distribution)Stat=med.2026.06 | 0.037 | 20.07 | 1.12 | |
| Original (Target Distribution)Stat=mean2026.06 | 0.041 | 20.37 | 1.121 | |
| DDPM BaselineStat=mean2026.06 | 0.043 | 21.54 | 1.148 | |
| DDPM BaselineStat=med.2026.06 | 0.043 | 20.35 | 1.152 |