ssim.py

import torch
import torch.nn.functional as F
from torch.autograd import Variable
import numpy as np
from math import exp

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

def _ssim(img1, img2, window, window_size, channel, size_average = True):
    mu1 = F.conv2d(img1, window, padding = window_size//2, groups = channel)
    mu2 = F.conv2d(img2, window, padding = window_size//2, groups = channel)
    mu1_sq = mu1.pow(2)
    mu2_sq = mu2.pow(2)
    mu1_mu2 = mu1*mu2

    sigma1_sq = F.conv2d(img1*img1, window, padding = window_size//2, groups = channel) - mu1_sq
    sigma2_sq = F.conv2d(img2*img2, window, padding = window_size//2, groups = channel) - mu2_sq
    sigma12 = F.conv2d(img1*img2, window, padding = window_size//2, groups = channel) - mu1_mu2

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

    ssim_map = ((2*mu1_mu2 + C1)*(2*sigma12 + C2))/((mu1_sq + mu2_sq + C1)*(sigma1_sq + sigma2_sq + C2))

    if size_average:
        return ssim_map.mean()
    else:
        return ssim_map.mean(1).mean(1).mean(1)

class SSIM(torch.nn.Module):
    def __init__(self, window_size = 11, size_average = True):
        super(SSIM, self).__init__()
        self.window_size = window_size
        self.size_average = size_average
        self.channel = 1
        self.window = create_window(window_size, self.channel)

    def forward(self, img1, img2):
        img1 = F.softmax(img1,dim=1)
        (_, channel, _, _) = img1.size()

        if channel == self.channel and self.window.data.type() == img1.data.type():
            window = self.window
        else:
            window = create_window(self.window_size, channel)
            
            if img1.is_cuda:
                window = window.cuda(img1.get_device())
            window = window.type_as(img1)
            
            self.window = window
            self.channel = channel


        return 1-_ssim(img1, img2, window, self.window_size, channel, self.size_average)



if __name__ =="__main__":
    loss = SSIM()
    a = torch.tensor([[[0.1,0.2,0.5],[0.7,0.5,0.9],[0.9,0.7,0.6]],[[0.6,0.2,0.5],[0.7,0.7,0.9],[0.9,0.9,0.6]],[[0.1,0.4,0.5],[0.9,0.1,0.9],[0.9,0.7,0.7]]])

    a = a.unsqueeze(dim=0)

    b = torch.tensor([[0,1,1],[1,2,0],[0,0,0]])
    b = b.unsqueeze(dim=0)

    loss(a,b)


    input = torch.ones(size = (3,48,48))
    kernel = torch.ones(size=(3,1,3,3))
    print(input.shape)
    F.conv2d(input,kernel)