How can I extract color feature of an image without extracting white color?

I = imread('coriander.png');
imshow(I);

R = I(:,:,1);
G = I(:,:,2);
B = I(:,:,3);

count = 0;

if R == 255 && G == 255 && B == 255
    count = count;
else
    count = count +1;
end

disp(count);

I don't want to extract white color and just want to count the number of pixel where is not white color.

Answer 1

[Walter Roberson]

count = nnz( R ~= 255 | G ~= 255 | B ~= 255 );

Or more simply,

count = nnz( ~all(I == 255, 3) );

[test test]

Thanks. By the way, how can I separate the values of RGB without calculating the white pixel values?

R = I(:,:,1);
G = I(:,:,2);
B = I(:,:,3);

just like to get the values as above, but without white pixel values.

[Image Analyst]

If you want a 2-D image, you can't because the white pixels must be in there because 2-D images must remain rectangular. If you want a 1-D vector, you can delete them

mask = R==255 & G==255 & B=255;
RwithoutWhite = R(~mask);
GwithoutWhite = G(~mask);
BwithoutWhite = B(~mask);

Again, those are vectors (lists of pixel values), NOT images. However you can reassign the white pixels to some other value if you want

R(mask) = 137; % Whatever value you want.

THIS will still be a 2-D image.