loss.py

import torch
from math import exp
import torch.nn.functional as F
import torch.nn as nn
from torch.nn.modules.utils import _pair, _quadruple
import torchvision


def gaussian(window_size, sigma):
    gauss = torch.Tensor([exp(-(x - window_size // 2) ** 2 / float(2 * sigma ** 2)) for x in range(window_size)])
    return gauss / gauss.sum()


def create_window(window_size, channel=1):
    _1D_window = gaussian(window_size, 1.5).unsqueeze(1)
    _2D_window = _1D_window.mm(_1D_window.t()).float().unsqueeze(0).unsqueeze(0)
    window = _2D_window.expand(channel, 1, window_size, window_size).contiguous()
    return window


def ssim(img1, img2, window_size=11, window=None, size_average=True, full=False, val_range=None):
    # Value range can be different from 255. Other common ranges are 1 (sigmoid) and 2 (tanh).
    if val_range is None:
        if torch.max(img1) > 128:
            max_val = 255
        else:
            max_val = 1

        if torch.min(img1) < -0.5:
            min_val = -1
        else:
            min_val = 0
        L = max_val - min_val
    else:
        L = val_range

    padd = 0
    (_, channel, height, width) = img1.size()
    if window is None:
        real_size = min(window_size, height, width)
        window = create_window(real_size, channel=channel).to(img1.device)

    mu1 = F.conv2d(img1, window, padding=padd, groups=channel)
    mu2 = F.conv2d(img2, window, padding=padd, groups=channel)

    mu1_sq = mu1.pow(2)
    mu2_sq = mu2.pow(2)
    mu1_mu2 = mu1 * mu2

    sigma1_sq = F.conv2d(img1 * img1, window, padding=padd, groups=channel) - mu1_sq
    sigma2_sq = F.conv2d(img2 * img2, window, padding=padd, groups=channel) - mu2_sq
    sigma12 = F.conv2d(img1 * img2, window, padding=padd, groups=channel) - mu1_mu2

    C1 = (0.01 * L) ** 2
    C2 = (0.03 * L) ** 2

    v1 = 2.0 * sigma12 + C2
    v2 = sigma1_sq + sigma2_sq + C2
    cs = torch.mean(v1 / v2)  # contrast sensitivity

    ssim_map = ((2 * mu1_mu2 + C1) * v1) / ((mu1_sq + mu2_sq + C1) * v2)

    if size_average:
        ret = ssim_map.mean()
    else:
        ret = ssim_map.mean(1).mean(1).mean(1)

    if full:
        return ret, cs
    return ret


class SSIM(torch.nn.Module):
    def __init__(self, window_size=11, size_average=True, val_range=None):
        super(SSIM, self).__init__()
        self.window_size = window_size
        self.size_average = size_average
        self.val_range = val_range

        # Assume 1 channel for SSIM
        self.channel = 1
        self.window = create_window(window_size)

    def forward(self, img1, img2):
        (_, channel, _, _) = img1.size()

        if channel == self.channel and self.window.dtype == img1.dtype:
            window = self.window
        else:
            window = create_window(self.window_size, channel).to(img1.device).type(img1.dtype)
            self.window = window
            self.channel = channel

        return ssim(img1, img2, window=window, window_size=self.window_size, size_average=self.size_average,
                    val_range=1)


class TVLoss(nn.Module):
    def __init__(self, TVLoss_weight=1):
        super(TVLoss, self).__init__()
        self.TVLoss_weight = TVLoss_weight

    def forward(self, x):
        batch_size = x.size()[0]
        h_x = x.size()[2]
        w_x = x.size()[3]
        count_h = self._tensor_size(x[:, :, 1:, :])
        count_w = self._tensor_size(x[:, :, :, 1:])
        h_tv = torch.pow((x[:, :, 1:, :] - x[:, :, :h_x - 1, :]), 2).sum()
        w_tv = torch.pow((x[:, :, :, 1:] - x[:, :, :, :w_x - 1]), 2).sum()
        return self.TVLoss_weight * 2 * (h_tv / count_h + w_tv / count_w) / batch_size

    def _tensor_size(self, t):
        return t.size()[1] * t.size()[2] * t.size()[3]


class VGGPerceptualLoss(torch.nn.Module):
    def __init__(self, weighting=1, resize=False):
        super(VGGPerceptualLoss, self).__init__()
        blocks = []
        blocks.append(torchvision.models.vgg16(pretrained=True).features[:4].eval())
        blocks.append(torchvision.models.vgg16(pretrained=True).features[4:9].eval())
        blocks.append(torchvision.models.vgg16(pretrained=True).features[9:16].eval())
        blocks.append(torchvision.models.vgg16(pretrained=True).features[16:23].eval())
        for bl in blocks:
            for p in bl:
                p.requires_grad = False
        self.blocks = torch.nn.ModuleList(blocks)
        self.transform = torch.nn.functional.interpolate
        self.mean = torch.nn.Parameter(torch.tensor([0.485, 0.456, 0.406]).view(1, 3, 1, 1))
        self.std = torch.nn.Parameter(torch.tensor([0.229, 0.224, 0.225]).view(1, 3, 1, 1))
        self.resize = resize
        self.weighting = weighting

    def forward(self, input, target, feature_layers=[0, 1, 2], style_layers=[2, 3]):
        if input.shape[1] != 3:
            input = input.repeat(1, 3, 1, 1)
            target = target.repeat(1, 3, 1, 1)
        input = (input - self.mean) / self.std
        target = (target - self.mean) / self.std
        if self.resize:
            input = self.transform(input, mode='bilinear', size=(224, 224), align_corners=False)
            target = self.transform(target, mode='bilinear', size=(224, 224), align_corners=False)
        loss = 0.0
        x = input
        y = target
        for i, block in enumerate(self.blocks):
            x = block(x)
            y = block(y)
            if i in feature_layers:
                loss += torch.nn.functional.l1_loss(x, y)
            if i in style_layers:
                act_x = x.reshape(x.shape[0], x.shape[1], -1)
                act_y = y.reshape(y.shape[0], y.shape[1], -1)
                gram_x = act_x @ act_x.permute(0, 2, 1)
                gram_y = act_y @ act_y.permute(0, 2, 1)
                loss += torch.nn.functional.l1_loss(gram_x, gram_y)
        return self.weighting * loss



#######################
#   Not use (only for experiment of dehazing)
#######################
class DCP(nn.Module):
    def __init__(self, kernel_size=3, stride=1, padding=0, same=True):
        super(DCP, self).__init__()
        self.k = _pair(kernel_size)
        self.stride = _pair(stride)
        self.padding = _quadruple(padding)
        self.same = same

    def _padding(self, x):
        if self.same:
            ih, iw = x.size()[2:]
            if ih % self.stride[0] == 0:
                ph = max(self.k[0] - self.stride[0], 0)
            else:
                ph = max(self.k[0] - (ih % self.stride[0]), 0)
            if iw % self.stride[1] == 0:
                pw = max(self.k[1] - self.stride[1], 0)
            else:
                pw = max(self.k[1] - (iw % self.stride[1]), 0)
            p1 = pw // 2
            pr = pw - p1
            pt = ph // 2
            pb = ph - pt
            padding = (p1, pr, pt, pb)
        else:
            padding = self.padding
        return padding

    def forward(self, x):
        x = F.pad(x, self._padding(x), mode='reflect')
        x = x.unfold(2, self.k[0], self.stride[0]).unfold(3, self.k[1], self.stride[1])
        x = x.contiguous().view(x.size()[:4] + (-1,))
        # print(x.median(dim=-1))
        x = x.min(dim=-1)[0].min(dim=1)[0]
        return x


class DCPLoss(nn.Module):
    def __init__(self, weighting=1, normalize=False):
        super(DCPLoss, self).__init__()
        self.weighting = weighting
        self.normalie = normalize
        self.dcp = DCP()

    def forward(self, x):
        x_dcp = self.dcp(x)
        out = torch.sum(x_dcp)
        return self.weighting * out / x.shape[0] / x.shape[2] / x.shape[3]