Maximum of Multidimensional Array using Flag Matrix

4 views (last 30 days)
Howdy all,
I have a question that should be relatively straight forward, but is giving me a world of trouble. I have two identically sized matrices: one with a range of values and a second matrix that will serve to identify subsets within the first matrix. I have broken down my larger matrices into a fundamental process I'd like to complete, namely: find the maximum of the first matrix only if the second matrix has a zero (no identifier) at that position.
The following simple script should allow for the maximum to be identified, and then I set the identifier matrix to one to eliminate it from the next maximum calculation:
A = magic(7);
B = zeros(7, 7);
Maximum = 49;
while (Maximum > 25)
[Maximum, Index] = max(A(B == 0));
[I, J] = ind2sub(size(A(B==0)), Index);
B(I,J) = 1;
disp([Data, I, J])
end
which works for the first few iterations (Maximum = 49-46), but then produces the wrong index and then removes 1 from the index each additional iteration until reaching a constant value. I'm not sure why MATLAB is behaving like it is, but I would really appreciate any input on this problem. Thanks!
  1 Comment
Stephen23
Stephen23 on 25 Feb 2017
@Everett625: your code throws errors: first undefined Data, and then when I comment out that line some problem with matrix sizes. Please edit your question and provide a working example if you want help.

Sign in to comment.

Answers (1)

Matt J
Matt J on 22 Feb 2017
Edited: Matt J on 22 Feb 2017
Are you sure you don't want all the maxima excluded in each pass (in case it is non-unique)? What about as follows?
while (Maximum > 25)
Maximum = max(A(B == 0));
B(A==Maximum & B==0) = 1;
end
  1 Comment
Everett625
Everett625 on 25 Feb 2017
This solution is fine for determining the maximum and setting the flag matrix correctly, but it seems there is something wrong with ind2sub. If I try to find the indices using either my original code or your simpler setup, the ind2sub command returns the wrong i and j values. And I need those to pass into a subroutine that performs a flood fill in my A matrix and marks zones in the B matrix. Am I missing something with ind2sub?

Sign in to comment.

Tags

Products

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!