Skip to main content

Table 6 Comparison of accuracy, recall, F1 score, and AUC of the methods on test data by the deep network based on four data augmentation methods

From: Automated chest screening based on a hybrid model of transfer learning and convolutional sparse denoising autoencoder

Loss function Data augmentation Precision Recall F1 AUC
Cross entropy Without augmentation 0.76 0.72 0.74 0.821
0.91 0.93 0.92
SMOTE 0.58 0.82 0.68 0.813
0.93 0.81 0.87
Positive augmentation 0.67 0.75 0.81 0.815
0.92 0.88 0.90
4× augmentation 0.69 0.77 0.72 0.827
0.92 0.89 0.91
Focal loss Without augmentation 0.76 0.73 0.74 0.828
0.91 0.93 0.92
SMOTE 0.52 0.82 0.64 0.79
0.93 0.76 0.84
Positive augmentation 0.74 0.72 0.73 0.817
0.91 0.92 0.91
4× augmentation 0.7 0.74 0.72 0.821
0.9 0.9 0.91