Creating a probability map for an image

I have an image ("img") and two arrays - x axis as the pixel intensity and y axis as the probability of a pixel with this intensity to have some property (x and y form a PDF). I want to create a matrix ("map") with the size of img where on each position there's the probability of the pixel in that position on img to have the property, means: map(p) = y(img(p)). I've attempted to use a container map but the performance is very bad and it takes approx. 3 seconds per image to execute (bellow - unique image values is x, Fpdf is y). Is there any better way to achieve this?

MF = containers.Map(unique(img,'sorted'),Fpdf);
FGprobMap = zeros(size(img));
for i=(1:size(img,1))
    for j=(1:size(img,2))
        pixB = MF(img(i,j));
        FGprobMap(i,j) = pixG;
    end
end