Skip to main content

Table 1 Accuracy assessment for two methods using precision, recall, F1-score and mAP (IOU = 50)

From: Identifying and mapping individual medicinal plant Lamiophlomis rotata at high elevations by using unmanned aerial vehicles and deep learning

Study sites

Method

Precision

Recall

F1

mAP

S1

ResNet-50-FPN

92.75 \(\pm\) 0.61

98.56 \(\pm\) 1.04

97.99 \(\pm\) 1.27

97.59 \(\pm\) 1.65

ResNet-101-FPN

93.46 \(\pm\) 0.71

98.90 \(\pm\) 1.21

98.61 \(\pm\) 1.40

98.32 \(\pm\) 1.60

S2

ResNet-50-FPN

83.89 \(\pm\) 0.95

90.43 \(\pm\) 1.36

87.61 \(\pm\) 1.42

84.98 \(\pm\) 1.57

ResNet-101-FPN

85.21 \(\pm\) 0.74

91.38 \(\pm\) 1.86

88.51 \(\pm\) 2.05

85.81 \(\pm\) 2.26

  1. The value is the mean \(\pm\) standard deviation of 10 experiments