Table 1. CNN hyperparameter configuration.

Parameter	Notation	CNN Architecture (C1)
Learning rate	α	0.0001
Loss function	l(x)	categorical cross entropy
Epoch	e	5
Batch size	bs	32
Optimizer	θ t	Adam
Kernel size/count	f/k	[3,3]
Convolution layers	conv	[2conv-2conv]
Activation function	∑wibi	Relu
Pooling layers	P	[(2,2), (3,3)]
Padding/Stride	d / s	same / (1,1)