Semantic Segmentation on Matterport3D (val)
57.1mIoUDoReMi
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| DoReMi2025.11 | 57.1 | 68.2 | 83.3 | |
| Sonata2025.11 | 56.1 | 65.9 | 82.3 | |
| G2P2026.01 | 55.9 | 66.4 | 82.2 | |
| PT v3status=reproduced2026.01 | 55.5 | 64.9 | 82.1 | |
| PTv32025.11 | 54.5 | 65 | 81.5 | |
| DoReMiProtocol=Zero-shot domain generalization2025.11 | 49.4 | 61.9 | 78.6 | |
| SonataProtocol=Zero-shot domain generalization2025.11 | 48.1 | 61 | 77.6 | |
| Point-MoEProtocol=Zero-shot domain generalization2025.11 | 41.8 | — | — | |
| Point-MoE-LParams=100M, Activated=60M, Precise Evaluator (None)=None, Joint indoor-outdoor training=true2025.05 | 40.7 | — | — | |
| OURSModal=RGB-D-N2023.08 | 39.26 | — | — | |
| OURSModal=RGB-D2023.08 | 39.19 | — | — | |
| OURSModal=RGB-N2023.08 | 38.91 | — | — | |
| PTv3-L*Params=97M, Activated=97M, Precise Evaluator (None)=None, Joint indoor-outdoor training=true, Training batch composition=Mix dataset batch2025.05 | 38.9 | — | — | |
| PTv3-L**Params=97M, Activated=97M, Precise Evaluator (None)=None, Joint indoor-outdoor training=true, Training batch composition=Mix dataset batch, Normalization type (layernorm)=Yes2025.05 | 37.1 | — | — | |
| HoHoNetModal=RGB-D2023.08 | 35.36 | — | — | |
| OURSModal=RGB2023.08 | 35.15 | — | — | |
| PanoFormerModal=RGB-D2023.08 | 33.99 | — | — | |
| Trans4PASS+Modal=RGB2023.08 | 33.43 | — | — | |
| HoHoNetModal=RGB2023.08 | 31.91 | — | — | |
| PTv3-LParams=97M, Activated=97M, Precise Evaluator (None)=None, Joint indoor-outdoor training=true2025.05 | 31.8 | — | — | |
| PanoFormerModal=RGB2023.08 | 30.04 | — | — | |
| PPT-LParams=98M, Activated=98M, Precise Evaluator (None)=None, Joint indoor-outdoor training=true2025.05 | 20.8 | — | — |