Table 11 Feature extraction and training time for DNN model based on transfer learning features.

From: Multiple model visual feature embedding and selection method for an efficient ocular disease classification

Algorithm

No. of images

Time to extract features per image

Time to extract features in total images (Min.)

Epochs training and validation DNN model

Per epoch time to train and validate DNN model (in minutes)

Total time to train and validate DNN model (in minutes)

EfficientNetB3

6392

0.2

21

100

0.0045

0.45

DenseNet201

6392

0.24

25.5

100

0.0045

0.45

InceptionResNetV2

6392

2.7

28.7

100

0.0045

0.45

Combined Model

  

75.2

100

0.006

0.6