|Re: [eigen] Code review for .resizeLike()|
[ Thread Index |
| More lists.tuxfamily.org/eigen Archives
- To: eigen@xxxxxxxxxxxxxxxxxxx
- Subject: Re: [eigen] Code review for .resizeLike()
- From: Keir Mierle <mierle@xxxxxxxxx>
- Date: Mon, 2 Feb 2009 07:11:52 -0800
- Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma; h=domainkey-signature:mime-version:received:in-reply-to:references :date:message-id:subject:from:to:content-type; bh=tb30vEe1dNFYnEAnFBT5OcGgHC1vq7mh9scQLfeIkt0=; b=wuxHNsRYd69Q4bBDoO1Evw33mx/ZvqvmdTj0GI6oPzsiQawp6YAMb6MPHQNdzSSm+h wH9HgVqFshFrOoeuN+jOWoXWcG84N7fwik8QQIWfkUdnNbmGCsn8a/X0RGmhggoZrzAM CakvftUJw2LP5NBjO8vTV3lVbUcDZx1LsxNas=
- Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:in-reply-to:references:date:message-id:subject:from:to :content-type; b=E0U6xppHGcX3doQ8HHo9nd7Ul+vPUBOMe/9DZAznYwbxP7EfgCl7Ykqk2khk948A0m t75jmMUU2Pv58nsURx+5JeAmzhYKbx5ONbGMaKjyFWDHfBFFBUQ4DGf/+icGGd9F5YuN nDqDnK1ekSH3c8mpzbvZcWR0YR3mBa13/LOPI=
On Mon, Feb 2, 2009 at 2:32 AM, Gael Guennebaud <gael.guennebaud@xxxxxxxxx>
you know, you should really apply for a kde svn account and feel free
to commit such changes:
for "Who supports this request?" you can put either me or Benoit.
I applied yesterday :)
Also I've seen you have an implementation of the Levenberg–Marquardt
algorithm and this is something I would love to see as a Eigen's
module. What do you think ? Compared to your current version I would
put the solver as a template parameter so that one can use LU, SVD, QR
or whatever other solvers, like the sparse ones...
Yes, I'd like to have it in Eigen too. The reason I haven't sent it in is that it's really basic at the moment and hasn't had much testing. I was planning on using it for 2-view bundle adjustment. Also, there is some debate about how to best structure the API.
I'll probably send it along in a couple of weeks. In the meantime, you can watch it evolve in libmv in src/libmv/numeric/levenberg_marquardt.h and function_derivative.h
On an aside: would you guys be interested in an automatic-differentiation module? Something like that would be really useful for me (vision is mostly optimization). See http://homepage.mac.com/sigfpe/paper.pdf