Table 4 The detailed parameters of the Fus2Net

1

Input_Layer

(299, 299, 3)

(299, 299, 3)

None

None

2

Conv2d_1

(299, 299, 3)

(149, 149, 32)

(3, 3)

(2, 2)

3

Conv2d_2

(149, 149, 3)

(147, 147, 32)

(3, 3)

(1, 1)

4

Conv2d_3

(147, 147, 3)

(147, 147, 64)

(3, 3)

(1, 1)

Block 1 module 1

5

Max_pooling2d

(147, 147, 64)

(73, 73, 64)

(3, 3)

(2, 2)

6

Conv2d_4

(147, 147, 64)

(73, 73, 96)

(3, 3)

(2, 2)

7

Concatenate_1

(73, 73, 64)

(73, 73, 160)

None

None

(73, 73, 96)

Block 1 module 2

8

Conv2d_5

(73, 73, 160)

(73, 73, 64)

(1, 1)

(1, 1)

9

Conv2d_6

(73, 73, 64)

(71, 71, 96)

(3, 3)

(1, 1)

10

Conv2d_7

(73, 73, 160)

(73, 73, 64)

(1, 1)

(1, 1)

11

Conv2d_8

(73, 73, 64)

(73, 73, 64)

(7, 1)

(1, 1)

12

Conv2d_9

(73, 73, 64)

(73, 73, 64)

(1, 7)

(1, 1)

13

Conv2d_10

(73, 73, 64)

(71, 71, 96)

(3, 3)

(1, 1)

14

Concatenate_2

(71, 71, 96)

(71, 71, 192)

None

None

(71, 71, 96)

Block 1 module 2

15

Average_pooling2d

(71, 71, 192)

(35, 35, 192)

(3, 3)

(2, 2)

16

Conv2d_11

(71, 71, 192)

(35, 35, 192)

(3, 3)

(2, 2)

17

Concatenate_3

(35, 35, 192)

(35, 35, 384)

None

None

(35, 35, 192)

Block 2

18

Conv2d_12

(35, 35, 384)

(35, 35, 32)

(1, 1)

(1, 1)

19

Conv2d_13

(35, 35, 384)

(35, 35, 32)

(1, 1)

(1, 1)

20

Conv2d_14

(35, 35, 32)

(35, 35, 32)

(3, 3)

(1, 1)

21

Conv2d_15

(35, 35, 384)

(35, 35, 32)

(1, 1)

(1, 1)

22

Conv2d_16

(35, 35, 32)

(35, 35, 48)

(3, 3)

(1, 1)

23

Conv2d_17

(35, 35, 48)

(35, 35, 64)

(3, 3)

(1, 1)

24

Concatenate_4

(35, 35, 32)

(35, 35, 128)

None

None

(35, 35, 32)

(35, 35, 64)

25

Conv2d_18

(35, 35, 128)

(35, 35, 384)

(1, 1)

(1, 1)

26

Add

(35, 35, 384)

(35, 35, 384)

None

None

(35, 35, 384)

BN + ReLU

27

Average_pooling2d

(35, 35, 384)

(4, 4, 384)

None

(1, 1)

Dropout

Softmax