Skip to main content

Table 6 Confusion matrix for Normal, CNV, DME, and Drusen obtained by Resnet50 in Experiment #5

From: Classification of age-related macular degeneration using convolutional-neural-network-based transfer learning

True labels

 
 

Normal

CNV

DME

Drusen

Precision

F1-score

Average recall and specificity

Predicted labels

 Normal

242

0

1

0

0.9959

0.9979

 

 CNV

0

240

1

4

0.9796

0.9856

 

 DME

0

2

240

0

0.9917

0.9917

 

 Drusen

0

0

0

238

1.0000

0.9917

 

Recall

1.0000

0.9917

0.9917

0.9835

  

0.9917

Specificity

0.9986

0.9931

0.9972

1.0000

  

0.9972

Average precision and F1-score

    

0.9918

0.9917

 

Accuracy

0.9917

     Â