Skip to main content

Table 4 Segmentation results on the validation set of the 8 synthetic rosebush models. 40 synthetic rosebush models were used to train the networks

From: Segmentation of structural parts of rosebush plants with 3D point-based deep learning methods

  PointNet PointNet++ DGCNN ShellNet PointCNN RIConv
Re Flower 64.85 99.48 99.21 99.64 99.71 97.29
Leaf 95,28 98.43 99.59 99.54 99.56 97.98
Stem 45.16 97.82 97.36 96.87 98.23 78.71
Pr Flower 88.19 99.78 99.68 99.59 99.59 98.57
Leaf 81.42 99.39 99.28 99.20 99.56 94.28
Stem 77.39 94.04 98.15 98.22 98.32 91.60
IoU Flower 59.67 99.26 98.90 99.24 99.30 95.94
Leaf 78.27 97.84 98.87 98.74 99.13 92.48
Stem 39.90 92.11 95.60 95.19 96.61 73.41
MIoU 59.28 96.40 97.79 97.73 98.35 87.28
Acc 81.82 98.50 99.15 99.10 99.36 94.57