Table 5 Evaluation of baseline methods of transfer learning for pest classification.

From: Multiple model visual feature embedding and selection method for an efficient pest classification supporting precision agriculture

Parameters (Top to Bottom)

Models (Left to right)

DenseNet201

EfficientNetB3

InceptionResNetV2

Accuracy

94.86

96.56

96.39

Validation accuracy

88.44

94.04

89.17

Precision

98.15

98.90

98.42

Validation precision

92.07

96.04

92.77

Recall

90.49

92.61

92.70

Validation recall

87.59

92.82

87.59

Loss

0.46

0.41

0.39

Validation loss

0.71

0.53

0.88