Table 4.

For 10 times repeated learning, the combination of parameters having the maximum recognition rate for each iteration.

Iteration No.	Sequence No.	Parameter	Avg. Of Accuracy
1	942,110	Conv. Layer 1-$KS$:21, $KC$:256 Conv. Layer 2-$KS$:25, $KC$:128 Conv. Layer 3-$KS$:29, $KC$:64 Dense Layer 1-$DLNC$:2048 Dense Layer 2-$DLNC$:1024	$acc({\mathrm{Param}}_{D_{opt}}^{942110})$ = 93.51%
2	898,926	Conv. Layer 1-$KS$:21, $KC$:128 Conv. Layer 2-$KS$:21, $KC$:256 Conv. Layer 3-$KS$:29, $KC$:128 Dense Layer 1-$DLNC$:2048 Dense Layer 2-$DLNC$:1024	$acc({\mathrm{Param}}_{D_{opt}}^{898926})$ = 93.48%
…
8	937,339	Conv. Layer 1-$KS$:21, $KC$:256 Conv. Layer 2-$KS$:21, $KC$:64 Conv. Layer 3-$KS$:29, $KC$:256 Dense Layer 1-$DLNC$:1024 Dense Layer 2-$DLNC$:2048	$acc({\mathrm{Param}}_{D_{opt}}^{937339})$ =93.76%(maximum)
…
10	1,098,106	Conv. Layer 1-$KS$:25, $KC$:128 Conv. Layer 2-$KS$:21, $KC$:128 Conv. Layer 3-$KS$:29, $KC$:64 Dense Layer 1-$DLNC$:1024 Dense Layer 2-$DLNC$:1024	$acc({\mathrm{Param}}_{D_{opt}}^{1098106})$ =93.58%
Average			93.57%