# Problem in calculation formula

1 view (last 30 days)
Syakira Akmal on 4 Apr 2017
Commented: Guillaume on 5 Apr 2017
Hi, this my problem. "Subscript indices must either be real positive integers or logicals. Error in tryMVM (line23) s=sum(y(X-i,Y-j)-mean)". X and Y is my point. How can i convert double to logical?

KSSV on 4 Apr 2017
Show us the full code......we cannot help unless full code is known.
Syakira Akmal on 5 Apr 2017
this is my coding..
imshow(a);
x = roipoly(a);
figure, imshow(x);
y = imresize(x,0.5);
figure, imshow(y);
imwrite(y,'s1.jpg');
%to set point
[X,Y] = meshgrid(1:1224,1:1224);
%to calculate mean image
m = mean2(b)
%to calculate standard deviation
sd = std2(c)
%mean-variance moving window operation
for i=1:1224
for j=1:1224
s=sum(y(X-i,Y-j)-m) %problem
end

Guillaume on 4 Apr 2017
The problem you have is not that your subscripts are not logical, it is that either they're not positive or they're not integers. I'm fairly certain, even with the very limited info you've provided, that converting to logical won't give you the result you want.
So,
• make sure that X, Y, i and j are integers
• make sure that X > i and Y > j
• and, you never know, make sure that you haven't got a variable called sum that stops you from using the sum function.
And on that last point, don't use mean as a variable name since it prevents you from using the mean function.

Syakira Akmal on 5 Apr 2017
this is my coding..
imshow(a);
x = roipoly(a);
figure, imshow(x);
y = imresize(x,0.5);
figure, imshow(y);
imwrite(y,'s1.jpg');
%to set point
[X,Y] = meshgrid(1:1224,1:1224);
%to calculate mean image
m = mean2(b)
%to calculate standard deviation
sd = std2(c)
%mean-variance moving window operation
for i=1:1224
for j=1:1224
s=sum(y(X-i,Y-j)-m) %problem
end
I still cannot solve the problem...
Guillaume on 5 Apr 2017
Well, yes that code is not going to work. X(1) is 1, so for any i, the expression X-i is always going to result in at least one value less than 1. When i is 1224, none of the values in X-i are valid indices.
In any case, nothing in that loop makes sense and I've no idea what you're trying to achieve. If you're trying to find the mean of the image over a moving window then you can use conv2:
%mean of image over sliding window of size mxn:
conv2(img, ones(m, n)/(m*n))