Novel View Synthesis on DTU 9-view
28.22PSNRTWINGS
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| TWINGSViews=92026.05 | 28.22 | 0.958 | 0.04 | 0.024 | — | |
| OursNumber of views=92026.06 | 27.99 | 0.954 | 0.052 | — | — | |
| DropGaussianViews=92026.05 | 27.75 | 0.954 | 0.042 | 0.026 | — | |
| DropGaussianNumber of views=92026.06 | 27.65 | 0.95 | 0.052 | — | — | |
| StereoGSvariant=asterisk2026.06 | 27.39 | 0.955 | 0.047 | — | — | |
| ICO-GSTraining Views=92026.03 | 27.19 | 0.953 | 0.045 | — | — | |
| CoR-GSTraining Views=92026.03 | 27.18 | 0.947 | 0.045 | — | — | |
| CoR-GSViews=92026.05 | 27.18 | 0.947 | 0.045 | 0.029 | — | |
| CoR-GS2026.06 | 27.18 | 0.947 | 0.045 | — | — | |
| 3DGSViews=92026.05 | 26.85 | 0.942 | 0.049 | 0.032 | — | |
| StereoGS2026.06 | 26.83 | 0.955 | 0.049 | — | — | |
| BinocularGSTraining Views=92026.03 | 26.7 | 0.947 | 0.052 | — | — | |
| Binocular3DGS2026.06 | 26.7 | 0.947 | 0.052 | — | — | |
| FSGSViews=92026.05 | 26.61 | 0.943 | 0.064 | 0.034 | — | |
| MVPGS2026.06 | 26.45 | 0.946 | 0.047 | — | — | |
| DNGaussianViews=92026.05 | 25.95 | 0.924 | 0.074 | 0.04 | — | |
| FreeNeRFTraining Views=92026.03 | 25.38 | 0.888 | 0.102 | — | — | |
| FreeNeRFViews=92026.05 | 25.38 | 0.888 | 0.096 | 0.046 | — | |
| FreeNeRFNumber of views=92026.06 | 25.38 | 0.888 | 0.102 | — | — | |
| FreeNeRF2026.06 | 25.38 | 0.888 | 0.102 | — | — | |
| MuRFNumber of views=92026.06 | 25.28 | 0.936 | 0.084 | — | — | |
| DiffusioNeRFSetting=Optimized per Scene2023.02 | 25.18 | 0.883 | — | 0.047 | 0.095 | |
| SparseNeRF2026.06 | 25.13 | 0.871 | 0.114 | — | — | |
| MixNeRFApproach=Regularization2023.02 | 25.03 | 0.879 | 0.065 | 0.042 | — | |
| RegNeRFApproach=Regularization2023.02 | 24.93 | 0.884 | 0.089 | 0.047 | — | |
| RegNeRFSetting=Optimized per Scene2023.02 | 24.93 | 0.884 | — | 0.047 | 0.089 | |
| RegNeRFTraining Views=92026.03 | 24.93 | 0.884 | 0.089 | — | — | |
| RegNeRFViews=92026.05 | 24.93 | 0.884 | 0.089 | 0.047 | — | |
| RegNeRF2026.06 | 24.93 | 0.884 | 0.089 | — | — | |
| ReconFusionNumber of views=92026.06 | 24.62 | 0.921 | 0.094 | — | — | |
| FSGSTraining Views=92026.03 | 24.33 | 0.911 | 0.106 | — | — | |
| FSGSNumber of views=92026.06 | 24.33 | 0.911 | 0.106 | — | — | |
| FSGS2026.06 | 24.33 | 0.911 | 0.106 | — | — | |
| DNGaussianTraining Views=92026.03 | 23.94 | 0.887 | 0.131 | — | — | |
| DNGaussianNumber of views=92026.06 | 23.94 | 0.887 | 0.131 | — | — | |
| DNGaussian2026.06 | 23.94 | 0.887 | 0.131 | — | — | |
| DietNeRFApproach=Regularization2023.02 | 23.83 | 0.823 | 0.173 | 0.068 | — | |
| DietNeRFSetting=Optimized per Scene2023.02 | 23.83 | 0.823 | — | 0.068 | 0.173 | |
| DietNeRFTraining Views=92026.03 | 23.83 | 0.823 | 0.173 | — | — | |
| DietNeRFViews=92026.05 | 23.83 | 0.823 | 0.173 | 0.068 | — | |
| mip-NeRFApproach=N/A2023.02 | 23.58 | 0.879 | 0.092 | 0.056 | — | |
| mip-NeRFSetting=Optimized per Scene2023.02 | 23.58 | 0.879 | — | 0.056 | 0.092 | |
| Mip-NeRFViews=92026.05 | 23.56 | 0.877 | 0.1 | 0.057 | — | |
| 3DGSTraining Views=92026.03 | 22.9 | 0.816 | 0.173 | — | — | |
| 3DGSNumber of views=92026.06 | 22.9 | 0.816 | 0.173 | — | — | |
| 3DGS2026.06 | 22.9 | 0.816 | 0.173 | — | — | |
| MVSNeRFApproach=Pre-training, fine-tuning=false2023.02 | 22.4 | 0.853 | 0.135 | 0.089 | — | |
| DS-NeRFOptimization Loss=KL divergence2021.07 | 22.3 | 0.81 | 0.24 | — | — | |
| MVSNeRFApproach=Pre-training, fine-tuning=true2023.02 | 22.22 | 0.853 | 0.135 | 0.088 | — | |
| MVSNERF ftSetting=DTU + ft per Scene2023.02 | 22.22 | 0.853 | — | 0.069 | 0.135 | |
| DS-NeRFOptimization Loss=MSE2021.07 | 22.2 | 0.77 | 0.26 | — | — | |
| NeRF2021.07 | 22.1 | 0.82 | 0.26 | — | — | |
| Geometric BaselineSetting=Optimized per Scene2023.02 | 22.01 | 0.853 | — | 0.067 | 0.121 | |
| PixelNeRFApproach=Pre-training, fine-tuning=true2023.02 | 21.83 | 0.781 | 0.203 | 0.09 | — | |
| PixelNeRF ftSetting=DTU + ft per Scene2023.02 | 21.83 | 0.781 | — | 0.09 | 0.203 | |
| pixelNeRF-DTUTraining=Pretrained on DTU2021.07 | 21.1 | 0.76 | 0.34 | — | — | |
| SRFApproach=Pre-training, fine-tuning=true2023.02 | 20.75 | 0.785 | 0.205 | 0.093 | — | |
| PixelNeRFApproach=Pre-training, fine-tuning=false2023.02 | 20.4 | 0.768 | 0.22 | 0.1 | — | |
| metaNeRF-DTUTraining=Pretrained on DTU2021.07 | 20.2 | 0.67 | 0.35 | — | — | |
| SRFApproach=Pre-training, fine-tuning=false2023.02 | 18.35 | 0.752 | 0.232 | 0.12 | — |