Table 3 The detection accuracy of DL-ADD with and without preprocessing.

From: Deep learning based atomic defect detection framework for two-dimensional materials

 

Origin

Preprocess

Impurity

Void

Impurity

Void

Recall

0.49

0.88

0.77

0.92

Precision

0.94

0.33

0.98

0.57

F1-Score

0.64

0.48

0.86

0.70

F2-Score

0.54

0.66

0.80

0.82