Novel View Synthesis on Tanks&Temples (PSNR, SSIM, LPIPS, #G/M)
23.75PSNRGaussianPOP
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| GaussianPOPTraining Iterations=40,0002026.02 | 23.75 | 85 | 18.1 | 0.18 | |
| GaussianPOPTraining Iterations=30,0002026.02 | 23.64 | 84.4 | 18.6 | 0.252 | |
| 3DGS2026.02 | 23.63 | 85 | 18 | 1.83 | |
| GaussianSpaTraining Iterations=40,0002026.02 | 23.61 | 84.8 | 18.3 | 0.185 | |
| LP-3DGS-R2026.02 | 23.6 | 84.2 | 18.8 | 1.244 | |
| MaskGaussianTraining Iterations=40,0002026.02 | 23.57 | 83.7 | 20 | 0.384 | |
| Mini-Splatting2026.02 | 23.45 | 84.1 | 18.6 | 0.319 | |
| CompGS2026.02 | 23.44 | 83.8 | 19.8 | 0.52 | |
| LP-3DGS-M2026.02 | 23.41 | 83.4 | 19.8 | 1.116 | |
| GaussianSpaTraining Iterations=30,0002026.02 | 23.4 | 84.1 | 18.8 | 0.295 | |
| EAGLES2026.02 | 23.37 | 84 | 20 | 0.65 | |
| Compact3DGS2026.02 | 23.32 | 83.1 | 20.1 | 0.836 | |
| LightGaussianTraining Iterations=40,0002026.02 | 23.1 | 80.4 | 22.8 | 0.251 |