|Re: [eigen] MatrixX multiply|
[ Thread Index |
| More lists.tuxfamily.org/eigen Archives
- To: eigen@xxxxxxxxxxxxxxxxxxx
- Subject: Re: [eigen] MatrixX multiply
- From: "Steven große Deters" <steven.grossedeters@xxxxxxxxx>
- Date: Thu, 15 Feb 2007 17:12:42 +0100
- Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=beta; h=received:message-id:date:from:to:subject:in-reply-to:mime-version:content-type:references; b=FdE/NJs1ayQxaEqEjUU3fPBecwg28YoYkiGdCVp7mKaHHnEBdXbkXwvfjrDMR7qOVvFUc5DCo09oxkhfj97vhtCee9uub/ha9eH+DO9pNx72MtTzZB3WVIESejAsAXpCst38OPbcBfLdxe1OL2JaYxwQDEwG4abOcUYjOa8IT0Y=
Thank you. Perhaps you could give me some advise so that I could try to do it on my own in the meantime ?
How could an entry-by-entry product of two (large) matrix's by accomplished ?
Would it be reasonable to access the single elements
e.g by row -> vector -> element and than multiply the elements. Or would that be to much of a performance hit ?
2007/2/15, Benoit Jacob <
> That is a little difficulty to explain. In short: I am calculating a
> Gaussian function for an image which describes the probability densities of
> the colour distribution. This is used for image segmentation.
> As a part of this calculation a covariance matrix (I don't know if that's
> the proper term in english) is needed. In order to calculate the covariance
> matrix I need the entry-by-entry product.
> (this explanation might not be totally correct - I am unfortunately not the
> mathematical master mind of this project)
OK, that sounds serious ;) I'll add entry-by-entry product as soon as
possible (which is perhaps not too soon as I'll be on a vacation between
Feb 17 and Feb 24).