| """ |
| Functions implementing custom NN layers |
| |
| Copyright (C) 2018, Matias Tassano <matias.tassano@parisdescartes.fr> |
| |
| This program is free software: you can use, modify and/or |
| redistribute it under the terms of the GNU General Public |
| License as published by the Free Software Foundation, either |
| version 3 of the License, or (at your option) any later |
| version. You should have received a copy of this license along |
| this program. If not, see <http://www.gnu.org/licenses/>. |
| """ |
| import torch |
| from torch.autograd import Function, Variable |
|
|
| def concatenate_input_noise_map(input, noise_sigma): |
| r"""Implements the first layer of FFDNet. This function returns a |
| torch.autograd.Variable composed of the concatenation of the downsampled |
| input image and the noise map. Each image of the batch of size CxHxW gets |
| converted to an array of size 4*CxH/2xW/2. Each of the pixels of the |
| non-overlapped 2x2 patches of the input image are placed in the new array |
| along the first dimension. |
| |
| Args: |
| input: batch containing CxHxW images |
| noise_sigma: the value of the pixels of the CxH/2xW/2 noise map |
| """ |
| |
| N, C, H, W = input.size() |
| dtype = input.type() |
| sca = 2 |
| sca2 = sca*sca |
| Cout = sca2*C |
| Hout = H//sca |
| Wout = W//sca |
| idxL = [[0, 0], [0, 1], [1, 0], [1, 1]] |
|
|
| |
| if 'cuda' in dtype: |
| downsampledfeatures = torch.cuda.FloatTensor(N, Cout, Hout, Wout).fill_(0) |
| else: |
| downsampledfeatures = torch.FloatTensor(N, Cout, Hout, Wout).fill_(0) |
|
|
| |
| noise_map = noise_sigma.view(N, 1, 1, 1).repeat(1, C, Hout, Wout) |
|
|
| |
| for idx in range(sca2): |
| downsampledfeatures[:, idx:Cout:sca2, :, :] = \ |
| input[:, :, idxL[idx][0]::sca, idxL[idx][1]::sca] |
|
|
| |
| return torch.cat((noise_map, downsampledfeatures), 1) |
|
|
| class UpSampleFeaturesFunction(Function): |
| r"""Extends PyTorch's modules by implementing a torch.autograd.Function. |
| This class implements the forward and backward methods of the last layer |
| of FFDNet. It basically performs the inverse of |
| concatenate_input_noise_map(): it converts each of the images of a |
| batch of size CxH/2xW/2 to images of size C/4xHxW |
| """ |
| @staticmethod |
| def forward(ctx, input): |
| N, Cin, Hin, Win = input.size() |
| dtype = input.type() |
| sca = 2 |
| sca2 = sca*sca |
| Cout = Cin//sca2 |
| Hout = Hin*sca |
| Wout = Win*sca |
| idxL = [[0, 0], [0, 1], [1, 0], [1, 1]] |
|
|
| assert (Cin%sca2 == 0), 'Invalid input dimensions: number of channels should be divisible by 4' |
|
|
| result = torch.zeros((N, Cout, Hout, Wout)).type(dtype) |
| for idx in range(sca2): |
| result[:, :, idxL[idx][0]::sca, idxL[idx][1]::sca] = input[:, idx:Cin:sca2, :, :] |
|
|
| return result |
|
|
| @staticmethod |
| def backward(ctx, grad_output): |
| N, Cg_out, Hg_out, Wg_out = grad_output.size() |
| dtype = grad_output.data.type() |
| sca = 2 |
| sca2 = sca*sca |
| Cg_in = sca2*Cg_out |
| Hg_in = Hg_out//sca |
| Wg_in = Wg_out//sca |
| idxL = [[0, 0], [0, 1], [1, 0], [1, 1]] |
|
|
| |
| grad_input = torch.zeros((N, Cg_in, Hg_in, Wg_in)).type(dtype) |
| |
| for idx in range(sca2): |
| grad_input[:, idx:Cg_in:sca2, :, :] = grad_output.data[:, :, idxL[idx][0]::sca, idxL[idx][1]::sca] |
|
|
| return Variable(grad_input) |
|
|
| |
| upsamplefeatures = UpSampleFeaturesFunction.apply |
|
|