Thanks Doug for your reply. In fact I need both. I do this mainly for computation (in volume); and display (in surface) is also cool.
The dots (e.g. A1, A2, ..., An, ..., AN) were originally defined in volume of the template (icbm152). This can be implemented as an Nx3 matrix S, where S(n, :) = [nx, ny, nz] indicates the location of An in the volume.
RAS here I mean such a coordinate system that, [0 0 0] is AC and the 3 dimensions are taking right, anterior, superior as positive, e.g An([nx0 ny0 nz0]) indicates a location nx0 mm right, ny0 mm anterior and nz0 mm superior of AC.
The above two systems can be easily transformed by a transformation matrix.
I wish to know the correspondent locations of An in individual subject(B), as Bn. e.g. if A1 is center of V1 in the template, I wish to find the center of V1 in the subject as B1.
So far, all the locations are represented in volume. However, I guess to normalise the individual brain to the template, freesurfer will extract the volumes to surfaces. I imagine the procedure will be like An (volume representation of dots in template) -> ASn (surface representation of dots in template) -> BSn(surface representation of dots in individual) -> Bn(volume representation of dots in individual). If this assumption is correct, I would like to use Bn for my computation and ASn for display later.
I hope I have made myself clearer.
best
Peng