Deep Transfer Learning Based Multiway Feature Pyramid Network for Object Detection in Images

Table 1

Architecture of EfficientNet.

Level x	Convolution function FX	Resolution HX × WX	No. of channels CX	No. of layers LX
1	Convol 3 × 3	224 × 224	32	1
2	k3 × 3, IRB1	112 × 112	16	1
3	k3 × 3, IRB6	112 × 112	24	2
4	k3 × 3, IRB6	56 × 56	40	2
5	k3 × 3, IRB6	28 × 28	80	3
6	k3 × 3, IRB6	14 × 14	112	3
7	k3 × 3, IRB6	14 × 14	192	4
8	k3 × 3, IRB6	7 × 7	320	1
9	Convol 1 × 1, pooling, FC	7 × 7	1280	1