Skipping 0 value in matrix computation wich lead to NaN.

Hello, I'm Derry.

first, forgive me for my bad english,

I'm currently working on my Thesis, but I've got something that bothering me. I tried to calculate an image Contrast Weighted Entropy (CWH) value, using this script :

img = imread('image.tif');
img = imresize(img, [32, 32]);
[M N] = size(img);
    numPix = M*N;
    [count, bins] = imhist(img);
    p = count/numPix;
    u=mean(count);
    val = (p-u).*p.* (log(p)/0.3010299957)*numPix;
    val2=sum(val);
    CWH=-1*val2;
    imshow (img);

but it give me a NaN value because it contain a 0 value on "image.tif" matrix.

the point is, is there any way to pass a value of 0 so that the calculation can be completed?

thank you in advance

Derry Suia

[Stephen23]

Note that you shoudl avoid using pi as a variable name, because this is a very important inbuilt constant pi. For the same reason you should never use the names size, cell, length, i, j, etc, etc.

[Derry Suia]

Hello Mr. Cobeldick,

thank you for your response, I did change "pi" as value name, but it still contain 0 value in matrix.

do you have any suggestion for my problem?

thanks

Derry Suia

Answer 1

[Thorsten]

Before you take the log, get rid of the 0's:

p = p(p ~= 0);

[Derry Suia]

wow, it works like a charm!

thank you for for your answer, kind Sir/Madam!

cheers!

Derry Suia