Table 5.

The CNN architecture used in this experiment.

Layers	Type	Filter Size	Neuron Numbers	Output Shape
Layer 1	Convolution	3 × 3	32	$31\times 31\times 32$
Layer 2	Convolution	3 × 3	64	$29\times 29\times 64$
Layer 3	Max pool	–	–	$14\times 14\times 64$
Layer 4	Dropout	–	–	$14\times 14\times 64$
Layer 5	Flatten	–	–	12544 × 1
Layer 6	Dense	–	128	128 × 1
Layer 7	Dropout	–	–	128 × 1
Layer 8	Dense	–	7	7 × 1