TABLE 4.

Detailed hyper-parameters of 3D-SENet model.

Layer name	Sub-layer name	Kernel size	Strides	Filters	Output size
MRI_images	N/A	N/A	N/A	N/A	121 × 145 × 121
Conv	N/A	3 × 3 × 3	1	64	121 × 145 × 121
Maxpooling	N/A	3 × 3 × 3	2	N/A	60 × 72 × 60
Se_block1	Block1	3 × 3 × 3	2	256	30 × 36 × 30
Se_block1	Block2	3 × 3 × 3	1	256	30 × 36 × 30
Se_block1	Block3	3 × 3 × 3	1	256	30 × 36 × 30
Se_block2	Block1	3 × 3 × 3	2	512	15 × 18 × 15
Se_block2	Block2	3 × 3 × 3	1	512	15 × 18 × 15
Se_block2	Block3	3 × 3 × 3	1	512	15 × 18 × 15
Se_block3	Block1	3 × 3 × 3	2	1024	8 × 9 × 8
Se_block3	Block2	3 × 3 × 3	1	1024	8 × 9 × 8
Se_block3	Block3	3 × 3 × 3	1	1024	8 × 9 × 8
Se_block3	Block4	3 × 3 × 3	1	1024	8 × 9 × 8
Avg_pool	N/A	8 × 9 × 8	1	1024	1 × 1 × 1024
Softmax	N/A	N/A	N/A	N/A	2