Table 4.

Training and validation accuracy comparison of the proposed CNN-SEEIB model and pre-trained models.

Model	Input shape	Training set accuracy	Validation set accuracy
VGG16	(224,224,3)	100.0%	95.29%
Resnet50	(224,224,3)	100.0%	99.65%
Efficient NetB0	(224,224,3)	100.0%	99.87%
DenseNet121	(224,224,3)	100.0%	99.83%
MobileNetV2	(224,224,3)	100.0%	99.78%
InceptionV3	(299,299,3)	100.0%	99.89%
Inception-ResnetV2	(224,224,3)	100.0%	99.83%
Xception	(224,224,3)	100.0%	99.83%
Proposed CNN-SEEIB	(224,224,3)	100.0%	99.89%