Semantic Occupancy Prediction on ReplicaOcc (test)
55.65IoUFreeOcc (rgbd)
Evaluation Results
| Method | Links | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| FreeOcc (rgbd)Learning Paradigm=Training-free, Annotation (Geo.)=✗, Annotation (Sem.)=✗, Input=RGB-D2026.04 | 55.65 | 17.8 | 8.6 | 44.33 | 2 | 20.76 | 16.31 | 34.9 | 24.48 | 20.9 | |
| FreeOcc (mono)Learning Paradigm=Training-free, Annotation (Geo.)=✗, Annotation (Sem.)=✗, Input=Monocular2026.04 | 46.81 | 19.38 | 8.88 | 38.23 | 21 | 15.11 | 12.14 | 25.07 | 16.46 | 16.93 | |
| EmbodiedOccLearning Paradigm=Fully supervised learning, Annotation (Geo.)=Occupancy, Annotation (Sem.)=✓2026.04 | 22.91 | 0 | 0 | 0 | 1 | 0 | 0 | 0.01 | 0.01 | 0 | |
| GaussTRLearning Paradigm=Self-supervised learning, Annotation (Geo.)=Poses, Annotation (Sem.)=✗2026.04 | 15.01 | 0 | 0 | 0.1 | 0 | 0 | 0 | 0 | 0 | 0.01 | |
| GaussianOccLearning Paradigm=Self-supervised learning, Annotation (Geo.)=Poses, Annotation (Sem.)=✗2026.04 | 8.71 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |