Discover MakerZone

MATLAB and Simulink resources for Arduino, LEGO, and Raspberry Pi

Learn more

Discover what MATLAB® can do for your career.

Opportunities for recent engineering grads.

Apply Today

problem in finding Mean value

Asked by Vennila on 6 Jan 2013

Hi,

    I want to find mean value of a mammogram only by choosing the values greater than zero. I used for loops with condition but it gives 255 as answer constantly. I don't know what's the mistake i have done.  

Code:

[x,y] = size(I);

s = 0;

for i = 1:x

    for j = 1:y
        if I(i,j) > 0 
             s = s + I(i,j);
        end
    end

end

disp('s = ');

disp(s);

0 Comments

Vennila

3 Answers

Answer by Walter Roberson on 6 Jan 2013
Accepted answer

Change to

s = s + double(I(i,j));

Question: when you are calculating the mean, are you going to be dividing by the number of values in I, or by the number of non-negative values?

Also are you sure that I will be two-dimensional and not 3 dimensional?

1 Comment

Vennila on 6 Jan 2013

Thank you walter. It works. I am going to divide the sum by no. of non-negative values.

Corrected Code:

[x,y] = size(I);

s = 0;

cnt = 0;

for i = 1:x

    for j = 1:y
        if I(i,j) > 0 
             s = s + double(I(i,j));
             cnt = cnt + 1;
        end
    end

end

m = s/cnt;

Walter Roberson
Answer by Image Analyst on 6 Jan 2013

Why are you doing two loops, which will make it slow? Why not just do:

nonZeroPixels = yourImage > 0;
meanValue = mean(yourImage(nonZeroPixels));

The way above is vectorized, faster, and very MATLABish.

0 Comments

Image Analyst
Answer by Jan Simon on 6 Jan 2013

Alternative method without loops:

s  = mean(double(I(I > 0)));

2 Comments

Walter Roberson on 6 Jan 2013

mean() works even without the double()

Jan Simon on 6 Jan 2013

Yes, Walter, because the underlying SUM uses the DOUBLE format as default, when 'native' is not specified. I would prefer to let functions reply the same type as the input as default, therefore I tend to cast more often than needed.

Jan Simon

Contact us