Vectorize 2 nested 'for' loops with 3D matrix indexing

Hello,

This is my piece of code:

%Mono is a huge matrix
a = size(mono,1);
b = size(mono,2); 
imA = zeros(a,b,4);
imB = zeros(a,b,4);

for i = 1:a
    for j = 1:b
        [imA(i,j,:),imB(i,j,:)] = encode(mono(i,j)); % encode is my function that accepts a scalar and returns two non-identical 1x4 vectors.
    end
end

Is there a way I could vectorize this?

Tried messing around with bsxfun without success.

Cheers!

[Akira Agata]

Is it possible to revise your encode function to accept 1D or 2D array ?

[seb leaf]

I could, but it would considerably change my outputs. How do you envision what the output would look like if my function accepts 1D arrays knowing that the current output is 2 vectors for scalar inputs?

If possible, I'd prefer to keep my encode function as it is, but if I absolutely need to revise my function I'll do it.

[Cyrus Tirband]

I think there's no way around changing your encode function.

Make it accept N x M arrays, and then have it output two NxMx4 matrices, where each 1x1x4 vector corresponds to the scalar inputs at the same indices.

Vectorising this code will have to happen in the encode function, not in this snippet you showed here.