Table 5 DSM Results
From: Automatic LIDAR building segmentation based on DGCNN and euclidean clustering
| No | Dataset | Number of buildings | Buildings detected | Buildings undetected | Accuracy | IOU |
|---|---|---|---|---|---|---|
| 1 | Dataset 1 | 32 | 17 | 15 | 53.12% | 0.46 |
| 2 | Dataset 2 | 35 | 15 | 20 | 42.85% | 0.34 |
| 3 | Dataset 3 | 61 | 31 | 30 | 50.81% | 0.42 |
| 4 | Dataset 4 | 36 | 22 | 14 | 61.11% | 0.54 |
| 5 | Dataset 5 | 62 | 40 | 22 | 64.51% | 0.53 |
| 6 | Dataset 6 | 75 | 47 | 28 | 62.67% | 0.51 |
| 7 | Dataset 7 | 103 | 60 | 43 | 58.25% | 0.49 |
| 8 | Dataset 8 | 54 | 33 | 21 | 61.11% | 0.53 |