Z translate 3D matrix data

I have a 3D matrix (512x512x600) full of data and a second 2D matrix that indicates the amount that each voxal needs to be translated by in the z-direction.

As a smaller example:

If I have 3D matrix M and want to translate each value in the (x,y) =(1,1) and (2,2) postion upward by one plane as defined by the values in matrix T= [1 0 0; 0 1 0; 0 0 0] while the rest of the points remain in the original planes

M = val(:,:,1) =
     1     2     3
     4     5     6
     7     8     9
val(:,:,2) =
   100   200   300
   400   500   600
   700   800   900
val(:,:,3) =
    10    20    30
    40    50    60
    70    80    90

After translation I would want something that looked like this:

M = val(:,:,1) =
     0     2     3
     4     0     6
     7     8     9
val(:,:,2) =
   1   200   300
   400   5   600
   700   800   900
val(:,:,3) =
    100    20    30
    40    500    60
    70    80    90
val(:,:,3) =
    10     0     0
    0     50     0
    0      0      0

Answer 1

[Matt J]

One way,

[m,n,p]=size(M);

[I,J,K]=ndgrid(1:m,1:n,1:p);

Knew=K+T;

result = accumarray([I(:),J(:),Knew(:)],M(:)),

[Brittney Gorman]

I have attached the specific files. When I try to use this method I get the error message:

Error using accumarray

First input SUBS must contain positive integer subscripts.

[Matt J]

Your T matrix contains non-integer values, which your initial post does not discuss and which my code is not prepared to deal with. Perhaps modify your original example to show the outcome you expect of a non-integer translation. And what about negative translations? Do you have a definition for those?