[Insight-users] multiplying a vnl_matrix with a vnl_vector

[michiel mentink]

solved it.

used vnl libraries instead:

  vnl_matrix<double> M(3, 4);
  vnl_vector<double> v(4);


  M[0][0] = 1; M[0][1] = 0; M[0][2] = 0; M[0][3] = 2;
  M[1][0] = 0; M[1][1] = 1; M[1][2] = 0; M[1][3] = 3;
  M[2][0] = 0; M[2][1] = 0; M[2][2] = 1; M[2][3] = 4;

  v[0] = 0;   v[1] = 0;  v[2] = 0;  v[3] = 1;

  std::cout << "v: " << std::endl << v << std::endl << std::endl;
  std::cout << "M: " << std::endl << M << std::endl;

  v = M*v;

  std::cout << "result: " << std::endl << v << std::endl << std::endl;


cheers, Michael



On Wed, Mar 31, 2010 at 11:56 AM, michiel mentink <
michael.mentink at st-hughs.ox.ac.uk> wrote:

> ok I found out that actual multiplication is no problem, my problem lies
> with storing the end result.
>
> This works:
>
>
>   typedef itk::Matrix<double,3,4> Matrix3by4Type;
>   typedef itk::Vector<double,4> Vector4Type;
>
>   Matrix3by4Type Translation;
>   Vector4Type FinalOrigin;
>
>   Translation[0][0] = 1; Translation[0][1] = 0; Translation[0][2] = 0;
> Translation[0][3] = 2;
>   Translation[1][0] = 0; Translation[1][1] = 1; Translation[1][2] = 0;
> Translation[1][3] = 3;
>   Translation[2][0] = 0; Translation[2][1] = 0; Translation[2][2] = 1;
> Translation[2][3] = 4;
>
>   FinalOrigin[0] = 0; FinalOrigin[1] = 0; FinalOrigin[2] = 0;
> FinalOrigin[3] = 1;
>
>   std::cout << "finalorigin: " << std::endl << FinalOrigin << std::endl;
>   std::cout << "translation: " << std::endl << Translation << std::endl;
>
>   std::cout << "p :   "   << FinalOrigin   << std::endl ;
>   std::cout << "A*p : " << Translation*FinalOrigin << std::endl ;
>
>
> // but this doesn't work:
>
>  FinalOrigin = A*p
>
>
> When I make the vector a 1by4 Matrix instead:
>
> FinalOrigin *= Translation;
>  compile error: no match for 'operator*=' in 'FinalOrigin *= Translation'
>
>  FinalOrigin *= Translation.GetTranspose();
>
> runtime error:
> vnl_matrix_fixed<T, num_rows, num_cols>::vnl_matrix_fixed(const
> vnl_matrix<T>&) [with T = double, unsigned int num_rows = 4u, unsigned int
> num_cols = 4u]: Assertion `rhs.rows() == num_rows && rhs.columns() ==
> num_cols' failed.
> (which makes sense because the the number of rows and colums of both the
> vector and matrix are not the same)
>
> So vector still seems to be the right way to go, I just can't figure out
> how to save it in a vector...
>
>
> Does anybody know how I save my result?
>
> cheers, Michael
>
>
>
>
>
> On Wed, Mar 31, 2010 at 10:21 AM, michiel mentink <
> michael.mentink at st-hughs.ox.ac.uk> wrote:
>
>>
>> I'd like to do the following calculation, a translate:
>>
>> A =
>> [ 1  0  0  2 ]
>> [ 0  1  0  3 ]
>> [ 0  0  1  4 ]
>>
>> B = [ x y z 1 ]
>>
>> FinalOrigin = A * B' = [x+2 y+3 z+4 ]
>>
>> For that, I have got the following code:
>>
>>
>>
>>   typedef itk::Matrix<double,3,4> Matrix3by4Type;
>>   typedef itk::Vector<double,4> Vector4Type;
>>
>>   Matrix3by4Type Translation;
>>   Vector4Type      FinalOrigin;
>>
>> FinalOrigin[0] = origin[0]; FinalOrigin[1] = origin[1]; FinalOrigin[2] =
>> origin[2]; FinalOrigin[3] = 1;
>> FinalOrigin = Translation * FinalOrigin;
>>
>>
>>
>> 1)
>> Is it possible to multiply an itk::vector with an itk::matrix in ITK?
>>
>> 2)
>> There is no transpose available for the vnl_vector type, that is
>> implemented in ITK.
>> Or is there?
>>
>>
>> I know that there are other methods in ITK to do translation, but for my
>> understanding, I'd like
>> to figure this out.
>>
>> cheers,
>>
>> Michael
>>
>>
>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.itk.org/pipermail/insight-users/attachments/20100331/cf7a98f2/attachment.htm>