Octave function 获取矩阵中的连续列组

Question

I am trying to find an efficient way of extracting groups of n consecutive columns in a matrix. Example:

A = [0, 1, 2, 3, 4; 0, 1, 2, 3, 4; 0, 1, 2, 3, 4]; 
n = 3;

should produce an output similar to this:

answer = cat(3, ...
                [0, 1, 2; 0, 1, 2; 0, 1, 2], ...
                [1, 2, 3; 1, 2, 3; 1, 2, 3], ...
                [2, 3, 4; 2, 3, 4; 2, 3, 4]);

I know this is possible using a for loop, such as the following code snippet:

answer = zeros([3, 3, 3]);
for i=1:3
  answer(:, :, i) = A(:, i:i+2);
endfor

However, I am trying to avoid using a for loop in this case - is there any possibility to vectorize this operation as well (using indexed expressions)?

Answer 1

Using just indexing

ind = reshape(1:size(A,1)*n, [], n) + reshape((0:size(A,2)-n)*size(A,1), 1, 1, []); 
result = A(ind);

The index ind is built using linear indexing and implicit expansion .

Using the Image Package / Image Processing Toolbox

result = reshape(im2col(A, [size(A,1) n], 'sliding'), size(A,1), n, []);

Most of the work here is done by the im2col function with the 'sliding' option.