Table 2.

CNN parameters employed.

Sl.No.	Layer	Kernel size/stride	No. of kernels	Size
1	Input image			160 x 160 x 1
2	Convolution layer 1 (C1)	3 x 3/1	16	158 x 158 x 16
3	Pooling layer 1 (P1)	2 x 2/2		79 x 79 x 16
4	Convolution layer 2 (C2)	3 x 3/1	32	39 x 39 x 32
5	Pooling layer 2 (P2)	2 x 2/2		20 x 20 x 32
6	Convolution layer 3 (C3)	3 x 3/2	64	10 x 10 x 64
7	Flatten			6,400
8	Fully connected layer	512		512
9	Soft max	2		2