[vtkusers] projections and transformations

[David Gobbi]

On Tue, Feb 16, 2010 at 6:56 AM, David Doria <daviddoria+vtk at gmail.com> wrote:
>> It divides by the "w" coordinate, which is the 4th: (x, y, z, w).
>> When you call TransformPoint(x,y,z) it creates a homogeneous
>> coordinate (x,y,z,w) with w=1, and then multiplies it by the 4x4
>> matrix.  Then it returns (x/w, y/w, z/w).
>>
>>   David G
>>
>
> Ah right, again because it is a 4x4 projection matrix. What do you
> think about adding vtkMatrix3x4 for typical "P = [KR | -KRC]" camera
> projection matrices?

The 4x4 matrix is the basis for the VTK transforms, so a vtkMatrix3x4
would have to be converted to a vtkMatrix4x4 before it could be used.
My opinion is that a 3x4 matrix would be kind of useless.  A person
can just construct a 4x4 matrix that does the same thing.

  David G