Table 3 Baseline performance of pre-trained models.

From: Transformer-enhanced deep ensemble for multi-class liver disease classification using computed tomography images

Model

Class

Accuracy

Precision

Recall

F1-Score

MCC

ResNet50V2

Liver Cirrhosis

0.82

0.56

1.00

0.72

0.773

 

Fatty Liver

 

1.00

0.67

0.80

 
 

Hcc

 

1.00

0.95

0.97

 

DenseNet121

Liver Cirrhosis

0.74

0.57

1.00

0.73

0.68

 

Fatty Liver

 

1.00

0.47

0.64

 
 

Hcc

 

0.77

1.00

0.87

 

MobileNetV2

Liver Cirrhosis

0.69

0.51

1.00

0.67

0.62

 

Fatty Liver

 

1.00

0.35

0.52

 
 

Hcc

 

0.75

1.00

0.86

 
Back to article page