Novel View Synthesis on ScanNet (novel view)
29.74PSNRSurfelNeRF
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| SurfelNeRFPer-scene optimization=true, Time=0.2s, MVS depth estimation=true2023.04 | 29.74 | 0.92 | 0.211 | |
| SurfelNeRFPer-scene optimization=true, Time=0.2s2023.04 | 29.58 | 0.919 | 0.215 | |
| PointNeRFPer-scene optimization=true, Time=30s2023.04 | 28.99 | 0.829 | 0.324 | |
| NeRFusionPer-scene optimization=true, Time=38s2023.04 | 26.49 | 0.915 | 0.209 | |
| NeRFingMVSPer-scene optimization=true2023.04 | 26.37 | 0.903 | 0.245 | |
| IBRNetPer-scene optimization=true2023.04 | 25.14 | 0.871 | 0.266 | |
| ADOPPer-scene optimization=true, Time=1s2023.04 | 25.01 | 0.807 | 0.272 | |
| Instant-NGPPer-scene optimization=true, Time=0.03s2023.04 | 23.23 | 0.714 | 0.459 | |
| OursSemantic encoder=CLIP-LSeg, Training views=122025.03 | 23.042 | 0.755 | 0.359 | |
| FSGSSemantic encoder=CLIP-LSeg, Training views=122025.03 | 21.875 | 0.73 | 0.386 | |
| CoR-GSSemantic encoder=CLIP-LSeg, Training views=122025.03 | 21.801 | 0.735 | 0.388 | |
| DNGaussianSemantic encoder=CLIP-LSeg, Training views=122025.03 | 20.231 | 0.706 | 0.451 | |
| Feature 3DGSSemantic encoder=CLIP-LSeg, Training views=122025.03 | 19.702 | 0.678 | 0.403 | |
| Gau-GroupingSemantic encoder=CLIP-LSeg, Training views=122025.03 | 19.189 | 0.682 | 0.419 | |
| 3DOVSSemantic encoder=CLIP-LSeg, Training views=122025.03 | 18.836 | 0.652 | 0.535 |