How to make a 3D matrix from 3 vectors?

3 vectors, a b c, to be used as factor or filter for a 3D matrix V. For 2D it's easy and elegant:

d=a'*b;
f=M.*d; % M 2D matrix

The 3D filter can be done this way:

f=V*0.0;
for n=1:n3
    f(:,:,n)=V(:,:,n)*c(n);
end

Is there a more elegant way for this? For example make a 3D matrix D from the 3 vector and then

f=D.*V

Thanks

Answer 1

[Matt J]

For filtering, it is probably not a good idea to convolve with d or D. Instead, you should filter separably. For example, in 2D, you could do,

   f = conv2(a,b,M);

For element-wise multiplication in 3D, I similarly don't think building D is the way to go, as it creates an unnecessary extra array. I would probably do,

f=bsxfun(@times,V,   a(:)*b(:).');
f=bsxfun(@times,f,   reshape(c,1,1,[]));

But, if you really insist on building D, you could do

D = bsxfun(@times,  a(:)*b(:).',   reshape(c,1,1,[]));

Or, you could do

d=a(:)*b(:).';

D=reshape(d(:)*c(:).',  size(V));

[John]

Thanks. That opens a few windows.