Getting an error: Error using .* Integers can only be combined with integers of the same class, or scalar doubles.

After running the following code I am getting this error in the pol2cart function

[srad,sang,S] = specxture(f);

function [ srad, sang, S] = specxture( f )

S = fftshift(fft2(f));
%imtool (S);
S = abs(S);
%imshow(S,[]);
[M, N] = size(S);
x0 =uint8(M/2+1);
%x0=int16(x0);
y0 =uint8(N/2+1);
%y0=int16(y0);
rmax = uint8((min(M, N)/2)-1);
%rmax1=int16(rmax);
srad = zeros(1, rmax);
srad(1) = S(x0, y0);
%imtool (S);
for r = 2:rmax
    [xc, yc] = halfcircle(r, x0, y0);
    srad(r) = sum(S(sub2ind(size(S), xc, yc)));
end
[xc, yc] = halfcircle(rmax, x0, y0);
sang = zeros(1, length(xc));
for a = 1:length(xc)
    [xr, yr] = radial(x0, y0, xc(a), yc(a));
    sang(a) = sum(S(sub2ind(size(S), xr, yr)));
end
S = mat2gray(log(1+S));
end



function [xc, yc] = halfcircle(r, x0, y0)

        theta = 91:270;
        theta = double(theta*pi/180);
        [xc, yc] = pol2cart(theta, r);
        xc = round(xc)'+x0;
        yc = round(yc)'+y0;

function [x,y,z] = pol2cart(th,r,z)
x = r.*cos(th);
y = r.*sin(th);

Answer 1

[Walter Roberson]

You define rmax as uint8 . Your for loop with rmax as upper bound then becomes uint8 . You are passing that r to halfcircle and passing that uint8 to pol2cart along with the vector of double for theta. The operation of pol2cart involves r.*cos(theta) . With theta being a vector of double this becomes uint8 .* vector of double . That operation is not defined.

The workaround is to loop the pol2cart so that you pass in only a scalar theta as combining uint8 and scalar double is defined . You would then get out the x and y calculated in uint8 that you seem to be looking for.

[D Aghor]

Thank you for the answer. Can you please suggest the changes in code to implement the solution.

[D Aghor]

Tried looping pol2cart. It worked. Thanks!

[Walter Roberson]

I admit to being puzzled as to why you are using uint8 operations; it makes your calculation more difficult for no obvious benefit. I would have suggested, for example, changing

x0 =uint8(M/2+1);

to

x0 = floor(M/2+1);

and likewise for the others, and then just gone ahead with double.

Answer 2

[D Aghor]

As a part of my base paper implementation I am using this ''specxture'' function as it is from an Image Processing Text book.