Re: [eigen] Adapting code from OpenCV (a seemingly much faster SVD)

[ Thread Index | Date Index | More Archives ]

License appears to be BSD like. License gurus, what say you?

On Sat, Apr 18, 2009 at 6:25 PM, Hauke Heibel
<hauke.heibel@xxxxxxxxxxxxxx> wrote:
> Hi,
> After performing some comparisons with Eigen's SVD (see attachment), I am
> wondering whether it might be of interest or possible at all to use the
> decomposition from OpenCV. I would volunteer :) to port it - maybe in a
> first step as a new SVD class in the 'disabled' folder that enables
> repeatable comparisons. This would obviously only make sense in case it is
> possible at all to use the code I have from the OpenCV lib. Actually, I
> ported the code before to our own hand-crafted linear algebra lib. But for
> Eigen, one definitely needs to take care of license issues. Since I have
> very little knowledge about these things, I am wondering whether somebody
> could comment on the attached license file - afaik, it is the same as when I
> copied the SVD code.
> The test results from the attached PDF were generated with double precision,
> dynamic size matrices of dimensions 25x25 up to 200x200 - each decomposition
> has been run 100 times. The graph contains the mean decomposition time in
> seconds for different matrix sizes (y-axis). The standard deviations in the
> legend are the deviations that occured while performing one and the same
> decomposition 100 times.
> Any comments would be welcome...
> Regards,
> Hauke

Rohit Garg

Senior Undergraduate
Department of Physics
Indian Institute of Technology

Mail converted by MHonArc 2.6.19+