Image Synthesis on NANO-TEM synthetic samples n=6656 benchmark leaderboard · SOTA2 Research

Image Synthesis on NANO-TEM synthetic samples n=6656

0.587MS-SSIMMS-SSIM VAE

Evaluation Results

Method
MS-SSIM VAE2026.06	0.587	0.407	18.45	0.014	0.409	—	—
VAE2026.06	0.587	0.407	18.45	0.0143	0.409	—	—
MS-SSIM VAE2026.06	0.579	0.397	18.63	0.015	0.417	170.96	0.147
VAE2026.06	0.579	0.397	18.63	0.0151	0.417	—	—
DCGAN2026.06	0.508	0.276	18.15	0.018	0.295	144.67	0.093
DCGAN2026.06	0.508	0.276	18.15	0.0177	0.295	—	—
patch-based progressive training methodology for DDPMs2026.06	0.481	0.27	16.99	0.022	0.277	47.32	0.034
Ours2026.06	0.481	0.27	16.99	0.0218	0.277	—	—
DCGAN2026.06	0.472	0.268	18.67	0.014	0.305	—	—
DCGAN2026.06	0.472	0.268	18.67	0.0136	0.305	—	—
patch-based progressive training methodology for DDPMs2026.06	0.414	0.211	16.89	0.021	0.266	—	—
Ours2026.06	0.414	0.211	16.89	0.0205	0.266	—	—
Baseline DDPM2026.06	0.412	0.199	16.34	0.025	0.229	68.52	0.067
Base2026.06	0.412	0.199	16.34	0.0248	0.229	—	—
Baseline DDPM2026.06	0.377	0.169	16.42	0.023	0.217	—	—
Base2026.06	0.377	0.169	16.42	0.0228	0.217	—	—
Base2026.06	—	—	—	—	—	68.52	0.067
DCGAN2026.06	—	—	—	—	—	144.67	0.093
Ours2026.06	—	—	—	—	—	47.32	0.034
VAE2026.06	—	—	—	—	—	170.96	0.147