3D Semantic Segmentation on ScanNet 19 classes
52.12mIoUEmbodiedSplat (FastSAM + OpenSeg)
Evaluation results
| Method | Links | ||
|---|---|---|---|
| EmbodiedSplat (FastSAM + OpenSeg)Inputs=RGB-D, Online / Offline=Online2026.03 | 52.12 | 75.66 | |
| OpenSceneInputs=Point Cloud, RGB-D, Online / Offline=Offline2026.03 | 50.71 | 72.75 | |
| EmbodiedSAMInputs=RGB-D, Online / Offline=Online2026.03 | 48.11 | 71.45 | |
| EmbodiedSplat (FastSAM + OpenSeg)Inputs=RGB, Online / Offline=Online2026.03 | 46.22 | 70.37 | |
| EmbodiedSplat (FastSAM + LSeg)Inputs=RGB, Online / Offline=Online2026.03 | 42.23 | 58.12 | |
| EmbodiedSplat (SAM + CLIP)Inputs=RGB, Online / Offline=Online2026.03 | 33.12 | 59.03 | |
| Occam’s LGSInputs=RGB, Online / Offline=Offline2026.03 | 30.49 | 57.91 | |
| LUDVIGInputs=RGB, Online / Offline=Offline2026.03 | 29.34 | 56.98 | |
| Dr. SplatInputs=RGB, Online / Offline=Offline2026.03 | 28.38 | 55.85 | |
| CF3Inputs=RGB, Online / Offline=Offline2026.03 | 26.34 | 52.43 |