Table 5.

Model final parameters.

Model	Batch Size	Number of Epochs	Hidden Layer Size	Dropout	Learning Rate	Activation	Optimizer	Kernel Size
VGG16	150	200	8–96 neurons	0.1	0.00001	ReLu sigmoid	Adam	2 × 2
VGG19	150	200	8–96 neurons	0.1	0. 0001	ReLu sigmoid	RMSprop	2 × 2
InceptionV3	200	300	8–96 neurons	0.1	0.0001	ReLu sigmoid	Nadam	2 × 2
ResNet50	100	200	8–96 neurons	0.1	0.001	ReLu sigmoid	Adamax	2 × 2
ResNet101	250	300	8–96 neurons	0.1	0.0001	ReLu sigmoid	Adam	2 × 2
GoogLeNet	50	150	8–96 neurons	0.1	0.0001	ReLu sigmoid	SGD	2 × 2
MobileNetV2	250	300	8–96 neurons	0.1	0.01	ReLu sigmoid	RMSprop	2 × 2
AlexNet	100	150	8–96 neurons	0.1	0.00001	ReLu sigmoid	Adadelta	2 × 2
EfficientNet B7	200	300	8–96 neurons	0.1	0.000001	ReLu sigmoid	Adamax	2 × 2
DenseNet121	200	350	8–96 neurons	0.1	0.00001	ReLu sigmoid	Adagrad	2 × 2
NFNet	150	250	8–96 neurons	0.1	0.0001	ReLu sigmoid	Adadelta	2 × 2
Modified MobileNetV2 (Proposed Method)	300	400	8–96 neurons	0.1	0.0000001	ReLu	RMSprop	2 × 2