Acquiring blocks of data

I have a 100x100 matrix (Matrix A) and ive been trying to get a matrix B. Where each cell in matrix B contains an array of all the value that surround the corresponding cell in Block A in terms of a block with size N (Lets say its 2 for now). The results should be matrix B which is 100x100 where each cell contain an array of the surrounding data points. I hope this makes sense.

I would appreciate any help

[Voss]

Maybe an example A that's 5x5 or so and the corresponding output B (still assuming N == 2, say) would clarify exactly what you have in mind.

[jonas]

What are you going to do with matrix B afterwards? Perhaps conv2 or blockproc functions could solve the problem without building matrix B.

Answer 1

[Vinai Datta Thatiparthi]

Hey Ahmed,

Firstly, since your output should be a collection of arrays of different values and dimensions, using cell arrays is the correct way to go about solving the problem. This code is a simplified version of what you're trying to do:

matIn = randi(5,5,5);   % The input matrix
                        % Insert the 100x100 matrix in your case
cellOut = cell(5,5);    % Cell array to hold the output

matRef = zeros(size(matIn));   

for i=1:numel(cellOut)
    matRef(:) = 0;
    matRef(i) = 1;
    % Use convolution to get the neighbors
    cellOut{i} = matIn(conv2(matRef,[1,1,1;1,0,1;1,1,1],'same')>0)';    
end

Finally, to get the neighbors of any element in matIn with the indices (i,j), simply use

cellOut{i,j}

Further, to echo @Jonas thoughts, consider using conv2 in your application directly to get what you want, instead of having to go through these steps.

Hope this helps!