variable section value give me some mean value but std is zero

bro i have the same coding in which the db variable section value give me some mean value but std is zero.and also during detection its detect same class (class 1) for all images.please any solution

    function [F]=FeatureStatistical(im)
  im = double(im(:));
  m=mean(mean(im)); 
  s=std(std(im));
  F=[m s];

[Walter Roberson]

That's odd, I was sure that I had answered this Question, but nothing shows up. Ah well, my answer was the same as dpb's

[dpb]

Musta' not gotten saved, Walter--no Answer when I looked first; I sort on "Unanswered" rather than look at other stuff....

Answer 1

[dpb]

You've double-compensated for the 2D image in

im = double(im(:));
m=mean(mean(im)); 
s=std(std(im));

im is now a vector; so mean(im) --> single value. mean(value)-->value so the problem doesn't show up here but similarly s=std(std(im)); is taking the std deviation of a single value whose deviation is, in fact, zero.

Rewrite the function as

im = double(im(:));
m=mean(im); 
s=std(im);

and all will be well...and just as a minor point since there's no further use in the function for the two temporaries, I'd write it as

F=[mean(im) std(im)];

eliminating them entirely.

[Image Analyst]

For 2D images, you can use mean2() and std2(), both in the Image Processing Toolbox.