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

Table 6 Segmentation results on 8 real rosebush models from ROSE-X data set with PointNet++

PointNet++		I	II	III	S	S + I	S + II	S + III
Re	Flower	69.98	68.79	81.24	97.02	75.13	73.70	85.39
	Leaf	98.02	98.28	97.67	28.72	98.10	98.71	98.71
	Stem	78.73	85.75	87.55	48.81	84.57	87.73	89.03
Pr	Flower	87.80	95.10	87.30	10.60	84.26	91.61	91.57
	Leaf	95.42	96.12	97.27	96.01	96.39	96.55	97.58
	Stem	83.73	83.89	86.20	78.09	89.47	90.48	92.52
IoU	Flower	63.78	66.44	72.72	10.57	65.88	69.04	79.17
	Leaf	93.62	94.53	95.07	28.38	94.62	95.35	96.36
	Stem	68.29	73.63	76.79	42.93	76.91	80.32	83.05
MIoU		75.23	78.20	81.53	27.29	79.14	81.57	86.19
Acc		93.70	94.63	95.28	36.15	94.82	95.58	96.60

ISSN: 1746-4811