Table 4 The optimum parameters value trained for this study.

Learning Rate (\(\alpha\))

0.0012

0.0016

0.0014

Batch Size (\(N\))

128

256

192

Number of Epochs

50

60

55

Weight Decay (\({\lambda }_{3}\))

0.0005

0.0003

0.0004

Momentum (\(\beta\))

0.9

0.92

0.94

Dropout Rate (\(p\))

0.2

0.25

0.22

Data Augmentation

Random cropping, horizontal flipping

Random rotation, color jitter

Random cropping, vertical flipping

Initialization Scheme

He initialization

Xavier initialization

He initialization

Cost Function Value (\({L}_{total}\))

0.607

0.511

0.617