svd and real representation of quaternions

greenspun.com : LUSENET : quaternions : One Thread

hello, my problem is that i'm using the real representation of quaternions to compute SVDQ but i really do not manage to get back the singular vector with it. do you have any idea about that? benoit PS thanks to cayley dickinson i can have back the singular vectors

-- benoit roue (benoit.roue@lis.inpg.fr), May 13, 2003

Answers

Hello Benoit:

When I work with quaternions, every number or matrix is a quaternion, so the expression:

A . X = B

is composed of three quaternions (this is not the general situation, but the one I work with). If one wants to determine X, then multiply on the left by the inverse of A. The inverse of A is easy to calculate: it is the transpose of A divided by its norm, so:

X = Transpose(A) . B/norm(A)

X will have the matrix form of a quaternion, not a singular vector.

doug quaternions.com

-- Doug Sweetser (sweetser@alum.mit.edu), May 13, 2003.


Moderation questions? read the FAQ