Cross validation: Loop with unbalanced length

Question

Kamu 2017-4-29

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/337928-cross-validation-loop-with-unbalanced-length

编辑： Kamu 2017-4-29

Hi all,

it might be a simple question but I would like to ask for a neat way how to write a loop for the following problem:

Let's say I have an array A=A(1), A(2).. A(n) and B=B(1), B(2).. B(m). I first want to do some calculations in a way that:

First iteration: A(1) with B(1), B(2).. B(m)
Second iteration: A(2) with B(1), B(2).. B(m)
till: A(n) with B(1), B(2).. B(m)

Explanation of the problem: I have 2 binary images, let us name them A and B. The centroid coordinates were obtained from A and I want to know if it is inside a given area in image B. The number of centroid and area are unbalanced (e.g. 10 centroids, and 6 area). Each centroid can be assigned to a specific polygon and delete those centroids that could not determine within any polygon.

My starting code:

numA = 10;
numB = 5;
for u=1:numA
    coordinate = imageA(u).Centroid;
    xCentroid = coordinate(:, 2);
    yCentroid = coordinate(:, 1);
      for v=1:numB
          area = imageB(v).PixelList;
          xArea = area(:, 2);
          yArea = area(:, 1);
          % Check if centroid is within area
          [in, out] = inpolygon(xCentroid, yCentroid, xArea, yArea);
      end
   end

1 个评论
显示 -1更早的评论隐藏 -1更早的评论

Stephen23 2017-4-29

编辑：Stephen23 2017-4-29

在 MATLAB Online 中打开

The simplest solution would be to put all of the "some operations" into a function, and then call it using bsxfun. It would require just one simple line of code:

bsxfun(fun,A(:),B(:).')

After all, repeated code should be put into functions anyway:

https://www.mathworks.com/matlabcentral/answers/228557-experts-of-matlab-how-did-you-learn-any-advice-for-beginner-intermediate-users

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Geoff Hayes 2017-4-29

1
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/337928-cross-validation-loop-with-unbalanced-length#answer_265018

在 MATLAB Online 中打开

Kamu - you could try something like

 N = 42;
 M = 84;
 A = randi(255,N,1);
 B = randi(255,M,1);
 for u=1:N
     a = A(u);
     for v=1:M
         b = B(v);
         % do some calculation with a and b
     end
 end

In the above, the outer loop iterates over each element of A and the inner loop over each element of B which should satisfy your requirements of fixing an element of A and using it with all elements of B.1

1 个评论
显示 -1更早的评论隐藏 -1更早的评论

Kamu 2017-4-29

编辑：Kamu 2017-4-29

在 MATLAB Online 中打开

Dear Geoff, thanks for your suggestion. However, it seems to be more complicated. Please find below my code:

numA = 10;
numB = 5;
for u=1:numA
    coordinate = imageA(u).Centroid;
    xCentroid = coordinate(:, 2);
    yCentroid = coordinate(:, 1);
      for v=1:numB
          area = imageB(v).PixelList;
          xArea = area(:, 2);
          yArea = area(:, 1);
          % Check if centroid is within area
          [in, out] = inpolygon(xCentroid, yCentroid, xArea, yArea);
      end
   end
    table = repmat(struct('IndexofA',[], 'Location', []), numA, 1);

So the idea is that I want to know if the Centroid is within a given polygon. The total number of centroid and polygon is unequal. The result should be stored in a structure element. The overall goal is to assign each centroid to a specific polygon and delete those centroids that could not determine within any polygon.

I hope this will make it clear and thanks again for your time and help.

请先登录，再进行评论。

Answer 2

Image Analyst 2017-4-29

1
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/337928-cross-validation-loop-with-unbalanced-length#answer_265043

在 MATLAB Online 中打开

I'll bet anything you got imageA(u) from doing this:

imageA = regionprops(labeledImage, 'Centroid');

Am I right? And you want to find out if the centroid is in the shape, like it's an O shape, or outside the shape, like a "C" shape. If you're willing to round the centroid to the nearest pixel, then you can use ismember or intersect(). Untested code:

imageA = regionprops(labeledImage, 'Centroid', 'PixelList);
for k = 1 : length(imageA)
  if ~isempty(intersect(imageA(k).Centroid, imageA(k).PixelList, 'rows'))
      % Centroid is one of the blob pixels.
  else
      % Centroid is not one of the blob pixels.
  end
end

1 个评论
显示 -1更早的评论隐藏 -1更早的评论

Kamu 2017-4-29

编辑：Kamu 2017-4-29

Dear Image Analyst,

thanks for your answer. Your bet was close to what I would like to do. Please allow me to explain it again: I have 2 binary images, let us name them A and B. The centroid coordinates were obtained from A and I want to know if it is inside a given area in image B. The number of centroid and area are unbalanced (e.g. 10 centroids, and 6 area).

请先登录，再进行评论。