Table 2.

Proposed deep learning approach.

Layer	Type	Maps	Neurons	Filter size
0	Input Image	3	71 × 71	—
1	Conv-1	90	70 × 70	2 × 2
2	MP-1	90	35 × 35	2 × 2
3	Conv-2	180	32 × 32	4 × 4
4	MP-2	180	16 × 16	2 × 2
5	Conv-3	360	14 × 14	3 × 3
6	MP-3	360	7 × 7	2 × 2
7	Conv-4	720	6 × 6	2 × 2
8	MP-4	720	3 × 3	2 × 2
9	Conv-5	1440	2 × 2	2 × 2
10	Decision layer	—	720	1 × 1
11	FC-1	—	100	1 × 1
12	FC-2	—	2	1 × 1