Table 2.

Network parameters

Pre-trained network	Number of layers	Classifier	Activation function or ELM	Number of hidden neurons of ELM	Mini-batch size	Output size
ResNet-50	179	ELM	Sigmoid	5000	32	3