DR-Net: A Novel Generative Adversarial Network for Single Image Deraining

Table 2

Layer parameters of the discriminator sub-network.

Layer Name	Parameters
Input	Rain image
Layer 1	Conv1. (3,3,24), stride=2, padding=1;
Layer 2	Conv2. (3, 3, 48), stride=2, padding=1; PReLU
Layer 3	Conv3. (3, 3, 96), stride=2, padding=1; PReLU
Layer 4	Conv4. (3, 3, 192), stride=2, padding=1; PReLU
Layer 5	Conv5. (3, 3, 384), stride=2, padding=1; PReLU
Layer 6	Conv6. (3,3,1), stride=2, padding=1; Sigmoid
Output	1 or 0