Table 1 Semantic segmentation results on the Oakland dataset
From: Pseudo-labelling-aided semantic segmentation on sparsely annotated 3D point clouds
Method | Overall accuracy (%) | Average F-score (%) | Per-class F-scores(%) | ||||
|---|---|---|---|---|---|---|---|
|  |  |  | Foliage | Wire | Pole | Ground | Façade |
Pointwise [1] | 79.8 | 49.9 | 82.2 | 4.3 | 4.3 | 91.4 | 67.5 |
CRF-reg [22] | 96.0 | 66.2 | 93.6 | 32.0 | 11.9 | 99.0 | 94.5 |
Seg-aided [17] | 96.6 | 68.4 | 93.7 | 46.5 | 8.7 | 99.5 | 93.9 |
Supervised baseline | 88.1 | 55.9 | 75.7 | 7.9 | 16.5 | 98.3 | 81.3 |
Ours no kdist | 91.2 | 62.7 | 82.2 | 10.3 | 36.8 | 98.4 | 85.9 |
Ours with kdist | 96.6 | 74.2 | 92.0 | 40.2 | 46.2 | 99.3 | 93.3 |