Get Haralick Features on masked Image

I have a grayscale (uint8) image and I want to get the Coocurrence matrix for it but I want graycomatrix to ignore certain pixels (according to a logical mask). I have this code:

if true
  % code
  gray_im = imread('test.bmp');
  mask_im = im2bw(imread('test_mask.bmp'));

    I = gray_im;
    I(~mask_im) = NaN; %make NaN the pixels that are not inside the mask. Doesn't work for uint8.

    [glcm,SI] = graycomatrix(I,'NumLevels',32,'Offset',[-2 2]); %[2 0] 
    stats =graycoprops(glcm  , {'contrast', 'homogeneity', 'energy'} );
  end

According to the documentation graycomatrix ignores NaN values but I can't put NaN values in a uint8 matrix (they get converted to zero and aren't ignored). I also tried to convert the image to double first but graycomatrix returned an unexpected result. Thanks in advance.

Answer 1

[Antonio Mendes]

Did you solve it?