metrics.py

# AUTOGENERATED! DO NOT EDIT! File to edit: 01_metrics.ipynb (unless otherwise specified).

__all__ = ['gaussian', 'create_window', 'SSIM', 'ssim', 'psnr', 'mutual_information', 'nmi', 'metric_fastai',
           'ssim_fastai', 'psnr_fastai', 'nmi_fastai']

# Cell
import torch
import torch.nn.functional as F
from torch.autograd import Variable
import numpy as np
from math import exp
from fastai2.vision.all import *

# Cell
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, data_range=255., padding=False, size_average = True):
    if padding:
        padding = window_size // 2

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

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

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

    C1 = (0.01*data_range)**2
    C2 = (0.03*data_range)**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):
    "Structural similarity index (SSIM) is a commonly used metric for CycleGAN experiments as it evaluates the preservation of content rather than style"
    def __init__(self, window_size = 11, data_range=255, padding = False, size_average = True):
        super(SSIM, self).__init__()
        self.window_size = window_size
        self.size_average = size_average
        self.data_range = data_range
        self.padding = padding
        self.channel = 1
        self.window = create_window(window_size, self.channel)

    def forward(self, img1, img2):
        (_, 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 _ssim(img1, img2, window, self.window_size, channel, self.data_range, padding, self.size_average)

def ssim(img1, img2, window_size = 11, data_range=255, padding = False, size_average = True):
    (_, channel, _, _) = img1.size()
    window = create_window(window_size, channel)

    if img1.is_cuda:
        window = window.cuda(img1.get_device())
    window = window.type_as(img1)

    return _ssim(img1, img2, window, window_size, channel, data_range, padding, size_average)


# Cell
def psnr(pred, targs, data_range=255):
    mse = F.mse_loss(pred, targs)
    return 20 * torch.log10(data_range / torch.sqrt(mse))

# Cell
def mutual_information(img1, img2, numBins = 20):
    # We compute the mutual information between img1 and img2,
    # which are assumed to be grayscale images (stored as numpy arrays)
    # numBins is a parameter that affects the mutual information score.

    hgram, x_edges, y_edges = np.histogram2d(img1.ravel(), img2.ravel(), bins = numBins)

    pxy = hgram/ float(np.sum(hgram))
    px = np.sum(pxy, axis = 1) # marginal for x over y
    py = np.sum(pxy, axis = 0) # marginal for y over x
    px_py = px[:, None] * py[None,:] # Broadcast to multiply marginals
    nzs = pxy > 0 # Only non-zero pxy values contribute to the sum

    return np.sum(pxy[nzs] * np.log(pxy[nzs] / px_py[nzs]))

def nmi(img1, img2, numBins=20):
    img1 = (img1[:,0,...]*0.2989 + img1[:,1,...]*0.5870 + img1[:,2,...]*0.1140).unsqueeze(1)
    img2 = (img2[:,0,...]*0.2989 + img2[:,1,...]*0.5870 + img2[:,2,...]*0.1140).unsqueeze(1)
    nmi = []
    for i in range(img1.shape[0]):
        nmi.append(mutual_information(img1.cpu().detach().numpy(),img2.cpu().detach().numpy()))
    return np.mean(nmi)


# Cell
def metric_fastai(xb,yb,_,metric):
    (real_A, real_B) = xb[0]
    fake_A, fake_B, idt_A, idt_B = yb
    return metric((real_A/2 + 0.5)*255,(fake_B/2 + 0.5)*255)

ssim_fastai = partial(metric_fastai, metric=ssim)
psnr_fastai = partial(metric_fastai, metric=psnr)
nmi_fastai  = partial(metric_fastai, metric=nmi)