Table 2 Segmentation results from three deep learning networks on the test set
Metric | Pointnet | Pointnet++  | PVCNN | |
|---|---|---|---|---|
Mean IoU (%) | 39.68 | 87.09 | 89.12 | |
Accuracy (%) | 70.28 | 95.57 | 96.19 | |
IoU (%) | Main stem | 41.43 | 84.43 | 89.84 |
Branch | 7.42 | 80.42 | 81.35 | |
Bolls | 70.18 | 96.44 | 96.16 | |
Precision (%) | Main stem | 58.16 | 90.38 | 91.82 |
Branch | 31.21 | 91.17 | 94.47 | |
Bolls | 76.27 | 98.02 | 97.62 | |
Recall (%) | Main stem | 64.75 | 92.24 | 97.69 |
Branch | 9.24 | 87.33 | 85.82 | |
Bolls | 90.73 | 98.32 | 98.45 | |
F1 score (%) | Main stem | 61.28 | 91.3 | 94.66 |
Branch | 14.26 | 89.21 | 89.94 | |
Bolls | 82.87 | 98.17 | 98.03 | |