Semantic Segmentation on Toronto-3D
81.13Mean IoU (mIoU)EyeNet
Evaluation results
| Method | Links | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| EyeNetRGB features=not used2023.01 | 81.13 | 94.63 | 96.98 | 65.02 | 97.83 | 93.51 | 86.77 | 84.86 | 94.02 | 30.01 | — | |
| RandLARGB features=not used2023.01 | 77.71 | 92.95 | 94.61 | 42.62 | 96.89 | 93.01 | 86.51 | 78.07 | 92.85 | 37.12 | — | |
| MappingConvSegRGB features=not used2023.01 | 77.57 | 93.17 | 95.02 | 39.27 | 96.77 | 93.32 | 86.37 | 79.11 | 89.81 | 40.89 | — | |
| global-local fusion frameworkvariant=Fusion+Geo2026.04 | 73.71 | — | 94.59 | 0 | 96 | 91.85 | 86.82 | 82.24 | 92.27 | 45.88 | — | |
| global-local fusion frameworkvariant=NoCrop+Proto+Focal2026.04 | 72.53 | — | 94.33 | 0 | 95.41 | 90.99 | 86.61 | 79.1 | 88.26 | 45.55 | — | |
| global-local fusion frameworkvariant=Dual-branch Fusion2026.04 | 72.51 | — | 94.59 | 0 | 95.8 | 92.28 | 86.77 | 81.91 | 91.34 | 37.41 | — | |
| PTv32026.04 | 70.86 | — | 94.58 | 0 | 95.38 | 92.73 | 86.19 | 79.83 | 89.73 | 28.46 | — | |
| MS-TGNetRGB features=not used2023.01 | 70.5 | 95.71 | 94.41 | 17.19 | 95.72 | 88.83 | 76.01 | 73.97 | 94.24 | 23.64 | — | |
| MinkUNet34C2026.04 | 69.56 | — | 94.48 | 0 | 95.34 | 89.53 | 83.28 | 77.57 | 89.64 | 26.61 | — | |
| KPConvRGB features=not used2023.01 | 69.11 | 95.39 | 94.62 | 0.06 | 96.07 | 91.51 | 87.68 | 81.56 | 85.66 | 15.72 | — | |
| SpUNet2026.04 | 68.87 | — | 94.62 | 0 | 95.9 | 91.35 | 82.38 | 76.25 | 86.17 | 24.3 | — | |
| PLOVIS2026.04 | 68.3 | — | — | — | — | — | — | — | — | — | 78.4 | |
| Rim et al.RGB features=not used2023.01 | 66.87 | 72.55 | 92.74 | 14.75 | 88.66 | 93.52 | 81.03 | 67.71 | 39.65 | 56.9 | — | |
| MS-PCNNRGB features=not used2023.01 | 65.89 | 90.03 | 93.84 | 3.83 | 93.46 | 82.59 | 67.8 | 71.95 | 91.12 | 22.5 | — | |
| Full fine-tuningFine-tuning strategy=Full fine-tuning2026.04 | 63.5 | — | — | — | — | — | — | — | — | — | 73.4 | |
| MLP probingFine-tuning strategy=MLP probing2026.04 | 63.4 | — | — | — | — | — | — | — | — | — | 72.2 | |
| Decoder probingFine-tuning strategy=Decoder probing2026.04 | 63.3 | — | — | — | — | — | — | — | — | — | 74.1 | |
| AAD-Net2026.04 | 63.1 | — | — | — | — | — | — | — | — | — | 71.4 | |
| ERDA2026.04 | 63.1 | — | — | — | — | — | — | — | — | — | 72.3 | |
| DG-Net2026.04 | 62.7 | — | — | — | — | — | — | — | — | — | 70.7 | |
| OctFormer2026.04 | 62.05 | — | 92.5 | 0 | 92.59 | 81.89 | 80.99 | 66.98 | 62.85 | 18.63 | — | |
| DGCNNRGB features=not used2023.01 | 61.79 | 94.24 | 93.88 | 0 | 91.25 | 80.39 | 62.4 | 62.32 | 88.26 | 15.81 | — | |
| TGNetRGB features=not used2023.01 | 61.34 | 94.08 | 93.54 | 0 | 90.83 | 81.57 | 65.26 | 62.98 | 88.73 | 7.85 | — | |
| MS-TGNet2020.03 | 60.96 | 91.69 | 90.89 | 18.78 | 92.18 | 80.62 | 69.36 | 71.22 | 51.05 | 13.59 | — | |
| KPFCNN2020.03 | 60.3 | 91.71 | 90.2 | 0 | 86.79 | 86.83 | 81.08 | 73.06 | 42.85 | 21.57 | — | |
| PointNet++RGB features=not used2023.01 | 59.47 | 92.56 | 92.9 | 0 | 86.13 | 82.15 | 60.96 | 62.81 | 76.41 | 14.43 | — | |
| TGNet2020.03 | 58.34 | 91.64 | 91.39 | 10.62 | 91.02 | 76.93 | 68.27 | 66.25 | 54.1 | 8.16 | — | |
| MS-PCNN2020.03 | 58.01 | 91.53 | 91.22 | 3.5 | 90.48 | 77.3 | 62.3 | 68.54 | 53.63 | 17.12 | — | |
| PointNet++2020.03 | 56.55 | 91.21 | 91.44 | 7.59 | 89.8 | 74 | 68.6 | 59.53 | 53.97 | 7.54 | — | |
| PointNet++mode=MSG (Multi-Scale Grouping)2020.03 | 53.12 | 90.58 | 90.67 | 0 | 86.68 | 75.78 | 56.2 | 60.89 | 44.51 | 10.19 | — | |
| Stratified Transformer2026.04 | 51.27 | — | 93.81 | 0 | 86.54 | 73.02 | 57.62 | 28.01 | 67.95 | 3.17 | — | |
| DGCNN2020.03 | 49.6 | 89 | 90.63 | 0.44 | 81.25 | 63.95 | 47.05 | 56.86 | 49.26 | 7.32 | — | |
| PTv22026.04 | 44.26 | — | 87.34 | 0 | 75.39 | 60.77 | 21.09 | 38.7 | 56.23 | 14.55 | — |