Image Reconstruction on Kodak (test)
59.52PSNRMiraGe
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| MiraGeNumber of Gaussians=100k2024.10 | 59.52 | 0.9999 | — | — | — | — | — | |
| MiraGeNumber of Gaussians=70k2024.10 | 57.41 | 0.9998 | — | — | — | — | — | |
| GaussianImageNumber of Gaussians=70k2024.10 | 44.08 | 0.9985 | — | — | — | — | — | |
| I-NGP2024.10 | 43.88 | 0.9976 | — | — | — | — | — | |
| NeuRBF2024.10 | 43.78 | 0.9964 | — | — | — | — | — | |
| 3DGS2024.10 | 43.69 | 0.9991 | — | — | — | — | — | |
| Fast 2DGSK=100k2025.12 | 43.01 | 0.9988 | 5 | — | — | — | — | |
| Fast 2DGSK=50k2025.12 | 41.85 | 0.9982 | 5 | — | 4.29 | — | — | |
| WIRE2024.10 | 41.47 | 0.9939 | — | — | — | — | — | |
| Instant-GIK=47,2462025.12 | 41.41 | 0.996 | 10 | — | 156.39 | — | — | |
| SIREN2024.10 | 40.83 | 0.996 | — | — | — | — | — | |
| Image-GS (g)K=100k2025.12 | 39.86 | 0.9949 | 30.64 | 508 | — | — | — | |
| GaussianImage (RS)K=50k2025.12 | 39.78 | 0.9971 | 13.66 | 2,459 | — | — | — | |
| Image-GS (s)K=100k2025.12 | 39.73 | 0.9949 | 27.86 | 509 | — | — | — | |
| GaussianImage (Cholesky)K=50k2025.12 | 39.36 | 0.9968 | 12.85 | 2,645 | — | — | — | |
| GaussianImage (Cholesky)K=100k2025.12 | 39.31 | 0.9952 | 14.8 | 2,257 | — | — | — | |
| Image-GS (g)K=50k2025.12 | 39.04 | 0.9939 | 27.97 | 771 | — | — | — | |
| GaussianImageNumber of Gaussians=100k*2024.10 | 38.93 | 0.9948 | — | — | — | — | — | |
| Image-GS (s)K=50k2025.12 | 38.65 | 0.9939 | 26.48 | 777 | — | — | — | |
| Fast 2DGSK budget=50,000, Fine-tuning time=1s2025.12 | 38.14 | — | — | — | — | — | — | |
| Instant-GIK budget=50,000, Fine-tuning time=1s2025.12 | 37.89 | — | — | — | — | — | — | |
| FM-FINERTrain Time (min)=9.242025.09 | 36.06 | 0.938 | — | — | — | — | — | |
| FM-SIRENTrain Time (min)=7.142025.09 | 35.62 | 0.933 | — | — | — | — | — | |
| GaussianImage (RS)K=100k2025.12 | 34.81 | 0.9863 | 15.08 | 2,211 | — | — | — | |
| Image-GS (s)K=10k2025.12 | 34.12 | 0.9835 | 22.91 | 1,475 | — | — | — | |
| Fast 2DGSK=10k2025.12 | 33.79 | 0.9865 | 5 | — | — | — | — | |
| Image-GS (g)K=10k2025.12 | 33.26 | 0.9777 | 27.76 | 1,340 | — | — | — | |
| Image-GS (G)K budget=50,000, Fine-tuning time=1s2025.12 | 31.12 | — | — | — | — | — | — | |
| Image-GS (S)K budget=50,000, Fine-tuning time=1s2025.12 | 30.49 | — | — | — | — | — | — | |
| GaussianImage (RS)K=10k2025.12 | 30.08 | 0.9684 | 13.26 | 2,639 | — | — | — | |
| FINERTrain Time (min)=9.242025.09 | 29.98 | 0.86 | — | — | — | — | — | |
| TUNERTrain Time (min)=4.152025.09 | 29.86 | 0.835 | — | — | — | — | — | |
| PETrain Time (min)=6.762025.09 | 29.71 | 0.837 | — | — | — | — | — | |
| GaussianImage (RS)K budget=50,000, Fine-tuning time=1s2025.12 | 29.2 | — | — | — | — | — | — | |
| GaussianImage (Cholesky)K budget=50,000, Fine-tuning time=1s2025.12 | 28.19 | — | — | — | — | — | — | |
| WIRETrain Time (min)=33.432025.09 | 27.52 | 0.747 | — | — | — | — | — | |
| FreShTrain Time (min)=10.222025.09 | 26.76 | 0.733 | — | — | — | — | — | |
| MIRETrain Time (min)=40.092025.09 | 26.48 | 0.719 | — | — | — | — | — | |
| SIRENTrain Time (min)=7.142025.09 | 25.98 | 0.696 | — | — | — | — | — | |
| GaussTrain Time (min)=10.132025.09 | 23.28 | 0.553 | — | — | — | — | — | |
| SPDERTrain Time (min)=22.502025.09 | 22.9 | 0.564 | — | — | — | — | — | |
| GaussianImage (Cholesky)K=10k2025.12 | 19.54 | 0.9685 | 12.68 | 2,852 | — | — | — | |
| CADCBitrate (bpp)=0.00762026.02 | — | — | — | — | — | — | 58.5 | |
| DLFBitrate (bpp)=0.00792026.02 | — | — | — | — | — | — | 12.5 | |
| StableCodecBitrate (bpp)=0.00722026.02 | — | — | — | — | — | — | 29 |