Scene Completion on NYU dataset (test)
75mIoUCleanerS
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| CleanerSInputs=RGB+TSDF, Backbone=Segformer-B2, Input depth condition=noisy depth2023.03 | 75 | 88 | 83.5 | |
| PVA-NetInput=D, Resolution=(60, 60)2021.12 | 74 | 91.1 | 79.7 | |
| CleanerSInputs=RGB+TSDF, Backbone=ResNet50, Input depth condition=noisy depth2023.03 | 73.1 | 89.9 | 79.6 | |
| IMENetInput=RGB+D, Resolution=(60, 60)2021.12 | 72.1 | 90 | 78.4 | |
| IMENetInputs=RGB+D, Input depth condition=noisy depth2023.03 | 72.1 | 90 | 78.4 | |
| FFNetInputs=RGB+D+TSDF, Input depth condition=noisy depth2023.03 | 71.8 | 89.3 | 78.5 | |
| 3D sketch-aware semantic scene completion frameworkResolution=(60, 60), Trained on=NYU, RGB-D based=true2020.03 | 71.3 | 85 | 81.6 | |
| 3D-SketchInput=RGB+D, Resolution=(60, 60)2021.12 | 71.3 | 85 | 81.6 | |
| 3D-SketchInputs=RGB+TSDF, Backbone=ResNet50, Input depth condition=noisy depth2023.03 | 71.3 | 85 | 81.6 | |
| CCPNetResolution=(240, 240), Trained on=NYU, RGB-D based=false2020.03 | 63.5 | 74.2 | 90.8 | |
| CCPNetInput=D, Resolution=(240, 240)2021.12 | 63.5 | 74.3 | 90.8 | |
| CCPNetInputs=TSDF, Input depth condition=noisy depth2023.03 | 63.5 | 74.2 | 90.8 | |
| ForkNetResolution=(80, 80), Trained on=NYU, RGB-D based=false2020.03 | 63.4 | — | — | |
| ForkNetfull model=true2019.09 | 63.4 | — | — | |
| ForkNetInput=D, Resolution=(80, 80)2021.12 | 63.4 | — | — | |
| ForkNetInputs=TSDF, Input depth condition=noisy depth2023.03 | 63.4 | — | — | |
| ForkNetcompletion branch=false2019.09 | 62.6 | — | — | |
| GRFNet2020.02 | 61.2 | 68.4 | 85.4 | |
| VVNetR-120Resolution=(120, 60), Trained on=NYU+SUNCG, RGB-D based=false2020.03 | 61.1 | 69.8 | 83.1 | |
| VVNet2019.09 | 61.1 | — | — | |
| ForkNetscene consistency=false2019.09 | 61.1 | — | — | |
| VVNetInput=D, Resolution=(120, 60)2021.12 | 61.1 | 69.8 | 83.1 | |
| DDR-SSC2020.02 | 61 | 71.5 | 80.8 | |
| DDRNetResolution=(240, 60), Trained on=NYU, RGB-D based=true2020.03 | 61 | 71.5 | 80.8 | |
| DDRNetInput=RGB+D, Resolution=(60, 60)2021.12 | 61 | 71.5 | 80.8 | |
| DDRNetInputs=RGB+D, Input depth condition=noisy depth2023.03 | 61 | 71.5 | 80.8 | |
| SATNet-TNetFuseResolution=(60, 60), Trained on=NYU+SUNCG, RGB-D based=true2020.03 | 60.6 | 67.3 | 85.8 | |
| SaTNet2019.09 | 60.6 | — | — | |
| SATNetInput=RGB+D, Resolution=(60, 60)2021.12 | 60.6 | 67.3 | 85.8 | |
| TS3DResolution=(240, 60), Trained on=NYU, RGB-D based=true2020.03 | 60 | — | — | |
| TS3DInput=RGB+D, Resolution=(240, 60)2021.12 | 60 | — | — | |
| AIC-NetInput=RGB+D, Resolution=(60, 60)2021.12 | 59.2 | 62.4 | 91.8 | |
| AIC-NetInputs=RGB+D, Input depth condition=noisy depth2023.03 | 59.2 | 62.4 | 91.8 | |
| SSCNet2019.09 | 56.6 | — | — | |
| EsscNet2020.02 | 56.2 | 71.9 | 71.9 | |
| ESSCNetResolution=(240, 60), Trained on=NYU, RGB-D based=false2020.03 | 56.2 | 71.9 | 71.9 | |
| ESSCNetInput=D, Resolution=(240, 60)2021.12 | 56.2 | 71.9 | 71.9 | |
| SSCNet2020.02 | 55.1 | 57 | 94.5 | |
| SSCNetResolution=(240, 60), Trained on=NYU, RGB-D based=false2020.03 | 55.1 | 57 | 94.5 | |
| SSCNetInput=D, Resolution=(240, 60)2021.12 | 55.1 | 57 | 94.5 | |
| SSCNetInputs=TSDF, Input depth condition=noisy depth2023.03 | 55.1 | 57 | 94.5 | |
| 3D-RecGAN2019.09 | 51.3 | — | — | |
| Geiger et al.2020.02 | 44.4 | 65.7 | 58 | |
| Geiger et al.Resolution=(240, 60), Trained on=NYU, RGB-D based=false2020.03 | 44.4 | 65.7 | 58 | |
| Geiger and Wang2019.09 | 44.4 | — | — | |
| IPF-SPCNetInputs=RGB+Point, Input depth condition=noisy depth2023.03 | 39 | 70.5 | 46.7 | |
| Lin et al.2020.02 | 36.4 | 58.5 | 49.9 | |
| Lin et al.Resolution=(240, 60), Trained on=NYU, RGB-D based=false2020.03 | 36.4 | 58.5 | 49.9 | |
| Lin et al.2019.09 | 36.4 | — | — | |
| SPCNetInput=D, Resolution=(240, 60)2021.12 | 36.3 | 72.1 | 42.2 |