Indoor Semantic Segmentation on ScanNet (val)
81.1mIoUUtonia
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| UtoniaProtocol=fine-tuning, Learnable Parameters=157.7M, Parameter Percentage=100%2026.03 | 81.1 | 89.1 | 93.3 | |
| ConcertoProtocol=fine-tuning, Learnable Parameters=124.8M, Parameter Percentage=100%2026.03 | 80.7 | 87.4 | 93.1 | |
| Volt-B#Params=94.3M, Training Setup=Jointly trained on multiple datasets2026.04 | 80.5 | — | — | |
| UtoniaProtocol=decoder tuning, Learnable Parameters=20.5M, Parameter Percentage=13%2026.03 | 80.3 | 88.3 | 92.9 | |
| Volt-S#Params=23.7M, Training Setup=Jointly trained on multiple datasets2026.04 | 80.2 | — | — | |
| DoReMiSource=-2025.11 | 80.1 | 87.3 | 93.1 | |
| ConcertoProtocol=decoder tuning, Learnable Parameters=16.3M, Parameter Percentage=13%2026.03 | 79.5 | 87.6 | 92.6 | |
| SonataProtocol=fine-tuning, Learnable Parameters=124.8M, Parameter Percentage=100%2026.03 | 79.4 | 86.1 | 92.5 | |
| SonataSource=CVPR 20252025.11 | 79.4 | 86.1 | 92.5 | |
| SonataProtocol=decoder tuning, Learnable Parameters=16.3M, Parameter Percentage=13%2026.03 | 79.1 | 86.6 | 92.7 | |
| PTv3/PPT#Params=124.8M, Training Setup=Jointly trained on multiple datasets2026.04 | 79.1 | — | — | |
| PPTProtocol=supervised, Learnable Parameters=124.8M, Parameter Percentage=100%2026.03 | 78.6 | 85.9 | 92.3 | |
| PPTSource=CVPR 20242025.11 | 78.6 | 85.9 | 92.3 | |
| MSCLearnable Parameters=124.8M, Parameter Percentage=100%2026.03 | 78.2 | 85.3 | 92.2 | |
| CDSegNetSource=CVPR 20252025.11 | 77.9 | 85.2 | 92.2 | |
| UtoniaProtocol=linear probing, Learnable Parameters=<0.2M, Parameter Percentage=<0.2%2026.03 | 77.7 | 87 | 91.9 | |
| PTv3Learnable Parameters=124.8M, Parameter Percentage=100%2026.03 | 77.6 | 85 | 92 | |
| PTv3Source=CVPR 20242025.11 | 77.6 | 85 | 92 | |
| PTv3#Params=46.1M, Training Setup=Standard2026.04 | 77.5 | — | — | |
| ConcertoProtocol=linear probing, Learnable Parameters=<0.2M, Parameter Percentage=<0.2%2026.03 | 77.3 | 86.6 | 91.7 | |
| Volt-S#Params=23.7M, Training Setup=Standard2026.04 | 77.2 | — | — | |
| Swin3D#Params=23.6M, Training Setup=Standard2026.04 | 76.4 | — | — | |
| OA-CNN#Params=51.5M, Training Setup=Standard2026.04 | 76.1 | — | — | |
| OctFormer#Params=44.0M, Training Setup=Standard2026.04 | 75.7 | — | — | |
| PTv2#Params=12.8M, Training Setup=Standard2026.04 | 75.4 | — | — | |
| ST#Params=18.8M, Training Setup=Standard2026.04 | 74.3 | — | — | |
| SonataProtocol=linear probing, Learnable Parameters=<0.2M, Parameter Percentage=<0.2%2026.03 | 72.5 | 83.1 | 89.7 | |
| MinkUNet#Params=39.2M, Training Setup=Standard2026.04 | 72.2 | — | — | |
| PointNeXt#Params=41.6M, Training Setup=Standard2026.04 | 71.5 | — | — | |
| PTv1#Params=11.4M, Training Setup=Standard2026.04 | 70.6 | — | — |