Whether the use of the function mahalanobis following is correct?

It's my first time to use mahalanobis for calculating & comparing distance of all projected trained images from the projected test image, whether the use of my coding is correct? Here is my code :

 mahal = [ ];
 for i=1 : size(eigenfaces,2)
    temp = (norm(projtestimg-projectimg(:,i)))^2;
    mahal = [mahal temp];
 end
 disp(mahal);
 [mahal_min recognized_index] = min(mahal);

I was a little doubt because I think that the results generated by the calculation is the same premise is that if I use euclidean. Please help me if I have been wrong to make code and please give the correct code for me