Image Decomposition Algorithm for Dual-Energy Computed Tomography via Fully Convolutional Network

Table 1

Detailed configuration of L-FCN/H-FCN.

Layer name	Kernel size	Stride	Pad	Output size
Input	—	—	—	65 × 65 × 1
Conv1	5 × 5	2	1	33 × 33 × 64
Conv2	5 × 5	2	1	17 × 17 × 128
Conv3	5 × 5	2	1	9 × 9 × 256
Conv4	9 × 9	1	0	1 × 1 × 512