Table 2.

Classification accuracy of different convolutional neuronal network (CNN) models during the optimization process.

A. CNN Models with Pretrained Weights on the ImageNet Dataset
CNN	VGG16				InceptionV3			InceptionResNetV2
Epochs, n	20		50		20	50		20		50
Training set	81%		82%		68%	70%		72%		74%
Validation set	81%		81%		59%	64%		62%		60%
B. CNN Models with Weights Trained on the Training Set
CNN	VGG16				InceptionV3			InceptionResNetV2
Epochs, n	20		50		20	50		20		50
Training set	88%		91%		83%	88%		87%		89%
Validation set	83%		86%		86%	85%		85%		84%
C. Different Image Input Sizes
Input size, px	128 × 128				256 × 256			395 × 395
Epochs, n	20				20			20
Training set	83%				95%			93%
Validation set	84%				89%			84%
D. Different Batch Sizes
Batch size, n		8		16			32		64
Epochs, n		20		20			20		20
Training set		84%		95%			94%		96%
Validation set		88%		89%			87%		89%
E. Different Dropout Rates
Dropout rate	0		0.1		0.2	0.3		0.4		0.5
Epochs, n	20		20		20	20		20		20
Training set	95%		89%		89%	88%		89%		88%
Validation set	89%		86%		84%	86%		86%		89%

CNN: Convolutional Neural Network.