Table 1.

Details of 2D-CNN Architecture.

Serial	Layer Type	Activations	Weights/Offset	Bias/Scale	Learnables
1	Image Input	81 × 128 × 1	-	-	0
2	Convolution	81 × 128 × 1	3 × 3 × 1 × 8	1 × 1 × 8	80
3	Batch Normalization	81 × 128 × 1	1 × 1 × 8	1 × 1 × 8	16
4	ReLU	81 × 128 × 1	-	-	0
5	Ma × Pooling	41 × 64 × 8	-	-	0
6	Convolution	41 × 64 × 16	3 × 3 × 8 × 16	1 × 1 × 16	1168
7	Batch Normalization	41 × 64 × 16	1 × 1 × 16	1 × 1 × 16	32
8	ReLU	41 × 64 × 16	-	-	0
9	Ma × Pooling	21 × 32 × 16	-	-	0
10	Convolution	21 × 32 × 16	3 × 3 × 16 × 16	1 × 1 × 16	2320
11	Batch Normalization	21 × 32 × 16	1 × 1 × 16	1 × 1 × 16	32
12	ReLU	21 × 32 × 16	-	-	0
13	Fully Connected	1 × 1 × 4	4 × 10,752	4 × 1	43,012
14	Softma×	1 × 1 × 4	-	-	0
15	Classification Output	-	-	-	0