Table 2. Characteristics of CCN used in the proposed framework.

CNN models	Year	Principal contribution	Parameters	Rate of error	Depth	Categorization	Reference
VGGNet	2014	Homogeneous topology Using fewer filters	138 M	ImageNet: 7.3	19	Spatial Exploitation	Simonyan & Zisserman (2014)
AlexNet	2012	More deep and broader than the LeNet Using Relu, drop and overlap Pooling NVIDIA GTX 580 GPU	60 M	ImageNet: 16.4	8	Spatial Exploitation	Krizhevsky, Sutskever & Hinton (2017)
ResNet	2016	Residual training Identity object tracking skip connection	25.6 M 1.7 M	ImageNet: 3.6 CIFAR-10: 6.43	152 110	Depth and Multipath	He et al. (2016)
GoogleNet	2015	Introduced principle of block Divide the idea of transformation and fusion	4 M	ImageNet: 6.7	22	Spatial Exploitation	Szegedy et al. (2016)