About Gaussian Meaning

[delldu]

Hi,
Thanks your good work in advance.
But I notices the followin code:

1. def gaussian_batch(self, dims):
   return torch.randn(tuple(dims)).to(self.device)

2. def test(self):
    Lshape = self.ref_L.shape
    input_dim = Lshape[1]
    self.input = self.real_H
    zshape = [Lshape[0], input_dim * (self.opt['scale']**2) - Lshape[1], Lshape[2], Lshape[3]]
    gaussian_scale = 1
    if self.test_opt and self.test_opt['gaussian_scale'] != None:
        gaussian_scale = self.test_opt['gaussian_scale']
    self.netG.eval()
    with torch.no_grad():
        self.forw_L = self.netG(x=self.input)[:, :3, :, :]
        self.forw_L = self.Quantization(self.forw_L)
        y_forw = torch.cat((self.forw_L, gaussian_scale * self.gaussian_batch(zshape)), dim=1)
        self.fake_H = self.netG(x=y_forw, rev=True)[:, :3, :, :]
    self.netG.train()
   

My question is:
What's the meaning with LR + Gaussian Noise to generate HR image ? Is it better than 0 or just for test ?

[JuZiSYJ]

I think the Gaussian noise is to sample the dropped high frequently information

[pkuxmq]

Yes, it is to sample from Gaussian noise and inversely recover a sample of possible lost information. Actually 0 is a sample from Gaussian noise, and under the full distribution training (i.e. IRN+), using 0 can also have a similar HR image.