Re: [eigen] refactoring fork: some news |

[ Thread Index |
Date Index
| More lists.tuxfamily.org/eigen Archives
]

I'm kind of having this need too. I'm writing algorithms where the matrices operate on a vector x, which is an image. As such, it sometimes need to be seen as a mxn array (not exactly a matrix), but most of the time it really is just a vector.
Examples of when it needs to be seen as an array is representation (transformation to a QImage) or computing the FFT (as a 2d object).
++
--
Thomas Capricelli <orzel@xxxxxxxxxxxxxxx>
http://www.freehackers.org/thomas
In data domenica 20 dicembre 2009 18:04:03, Manuel Yguel ha scritto:
: > Hi,
> I am practically sure that it is off the point however my concern is
> that in some applications, it is necessary to view a matrix of size
> nxm as a vector 1xnm or pxq with nm = pq.
> I wonder if this is the good moment to raise my case, but is this
> possible to consider having a matrix() function that takes two
> parameters that are compatible with the number of coefficients in the
> array: matrix(int rows, int cols ) with rows*cols = size()
> such that we can have for a 3x4 matrix, a 12x1 vector or a 2x6 matrix
> or a 4x3 matrix etc... assuming that the coefficients are given by the
> storage order row major or column major.
>
> - sincerely,
>
> Manuel
>
>
>