Table 3 The average test classification accuracy.

From: Classification and Morphological Analysis of Vector Mosquitoes using Deep Convolutional Neural Networks

DCNN model

Augmentation

Fine-tuning

LR

Accuracy (%)

VGG-16

  

5e-6

38.96

 

âś“

5e-6

56.76

âś“

 

5e-6

91.15

âś“

âś“

5e-6

97.19

ResNet-50

  

5e-3

57.47

 

âś“

1e-3

57.74

âś“

 

5e-3

93.45

âś“

âś“

7.5e-3

96.86

SqueezeNet

  

3e-5

47.42

 

âś“

3e-5

62.81

âś“

 

3e-5

78.58

âś“

âś“

3e-5

90.71

  1. Each model is trained with four settings of dynamic data-augmentation and fine-tuning. When a model is not fine-tuned, it is trained from scratch with random initialization.
Back to article page