Skip to main content

Table 6 The average precisions (AP) by class for EffDet V1 and V2 at an IoU threshold of 0.5 on the validation set

From: Automated corrosion detection in Oddy test coupons using convolutional neural networks

AP

EffDet V1 (%)

EffDet V2 (%)

Ag-P

96.39

72.94

Ag-T

94.18

75.31

Ag-U

95.41

76.30

Cu-P

96.30

60.37

Cu-T

87.20

58.96

Cu-U

89.72

72.65

Pb-P

86.96

58.03

Pb-T

93.69

72.60

Pb-U

94.36

76.97

mAP

92.69

69.35

  1. The mAP values differ slightly from Table 5 because they were calculated using another mAP package [53]