How to find euclidean distances between cell entries of two RGB matrices?

I have two RGB images A and B. Image A is 47x47x3 and image B is 1x456x3. I need to find the euclidean distances between all the entries of image A and all entries of image B. Then as a second goal I need to find for each cell in image A which cell in image B its closest to in euclidean distance. Does anyone has any suggestions? The things I tried have unfortunately failed.

[Jon]

It sounds like you want to do some kind of pixel to pixel comparison between image A and image B. I don't understand how you are defining this when image A has 47x47 = 2209 pixels and image B has 1x456 = 456 pixels. Besides the number of pixels being different (so how can you compare pixel to pixel) the dimensions are also different, one is a square matrix the other a column, so even if the number of pixels agreed which ones in the square image would correspond to which ones in the column image?

[Borys Bulka]

I am a bit confused myself how to exactly state what my goal is. What I try to say is that for each cell in RGB image A, I want to know what the euclidean distance is with each cell in RGB image B. Does that make any sense?

[Borys Bulka]

Maybe this will make more sense. In here I have the two images A and B: https://www.mathworks.com/matlabcentral/answers/644365-how-to-transform-a-rgb-matrix-to-a-weighted-adjacency-matrix?s_tid=prof_contriblnk.

A is a 47x47x3 RGB matrix. For each cell I want to know which cell in RGB matrix it is closest to in euclidean space. I am stuck on this problem for a whole week and I constantly cannot solve or find a way to do so. I am bit clueless on how to tackle it.

Answer 1

[Jan]

A = rand(47, 47, 3);
B = rand(1, 456 3);

AB = reshape(A, 47*47, 1, 3) - B;
Dist = vecnom(AB, 2, 3);

And now you want to find the minimal values in each column. Afterwards you can use ind2sub to convert the linear indices back to the indices of A.

Another option is a simple loop:

Result = zeros(size(A));
for row = 1:47
    for col = 1:47
        dist             = vecnorm(A(col, row, :) - B, 2, 3);
        [~, index]       = min(dist);
        Result(col, row) = index;
    end
end

[Borys Bulka]

Thank you a lot for your answer and effort, but somehow I fail to incorporate it and interpret the results correctly. I don't think any more answers will help since I have tried this in multiple directions ways with multiple answers and keep failing it.

[Jan]

I cannopt guess, which problem you have. So please post the relevant part of your code and ask a specific question.