Re: meaning of matrix operation A(B)

[Joao Rodrigues]

On 10/31/2014 07:17 AM, kankan wrote:
> Dear all,
>
> I have two matrix A and B. Size of A is 512x1 and size of B is 132x108.
> There is an operation in octave: A(B). Output of this operation is a matrix
> 132x108. What is the name of this operation and what does it is doing.
> Actualy I want to know the algorithm of this operation and want to use in
> fortran.
> The source code of applylut function of image-processing package of octave
> used this operation (at line no 42 of applylut.m).
> The command is : A=LUT(filter2(w,BW)+1);
> Here LUT is a matrix of dimension 512x1 and filter2(w,BW) is 132x108. I want
> to know how A has been calculated.
> Please help me.
> Thanks in advance.
>
> Regards,
> Kankan
what A(B) does is to return the elements of A indexed in B

octave:1> A = [1.1 2.2 3.3]'; B = [2 3 1]'; [A, B, A(B)]
ans =

   1.1000   2.0000   2.2000
   2.2000   3.0000   3.3000
   3.3000   1.0000   1.1000

octave:2> B = [2 3 1 4]'; [B, A(B)]
error: A(I): index out of bounds; value 4 out of bound 3
octave:2> B = [2 3 1 3]'; [B, A(B)]
ans =

   2.0000   2.2000
   3.0000   3.3000
   1.0000   1.1000
   3.0000   3.3000

octave:3> B = [2 3; 1 3]'; [B, A(B)]
ans =

   2.0000   1.0000   2.2000   1.1000
   3.0000   3.0000   3.3000   3.3000

Now if A is a matrix:

octave:4> A = [1.1 2.2 3.3; 4.4 5.5 6.6; 7.7 8.8 9.9]'
A =

   1.1000   4.4000   7.7000
   2.2000   5.5000   8.8000
   3.3000   6.6000   9.9000

Then A(B) returns the A values in indices B by reading A as a vector of 
concatenated columns:

octave:5> B = [2 4; 6 8]'; [B, A(B)]
ans =

   2.0000   6.0000   2.2000   6.6000
   4.0000   8.0000   4.4000   8.8000

You can also extract a sub-matrix from a matrix by specifying two index 
vectors (row and column);

octave:6> B = [2 3]'; C = [1 3]'; A(B,C)
ans =

   2.2000   8.8000
   3.3000   9.9000

I hope this helps.