Table 2 The result of K-fold cross-validation

From: Integrating deep learning and machine learning for ceramic artifact classification and market value prediction

K-fold cross-validation result

mAP@50 (%)

Recall (%)

Precision (%)

F1 score (%)

1st discount

69

90

99

66

2nd discount

69

92

99

62

3rd discount

69

91

98

65

4th discount

70

92

99

63

5th discount

68

90

99

65

  1. Best-F1 is the maximal value along the confidence sweep.