Newsgroups: sci.image.processing
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!news.mathworks.com!udel!gatech!howland.reston.ans.net!agate!usenet.ins.cwru.edu!bme.ri.ccf.org!neth
From: neth@bme.ri.ccf.org (Don Neth)
Subject: principal axis determination
Message-ID: <NETH.95Feb3120245@porky.bme.ri.ccf.org>
Sender: usenet@bme.ri.ccf.org
Nntp-Posting-Host: porky
Organization: Biomedical Engineering
Date: Fri, 3 Feb 1995 17:02:48 GMT
Lines: 13

I am attempting to align two 3D data sets using a principal axis
transformation approach.  I am able to determine the centroids and
moments of each volume.  Unfortuantely, determining the direction
cosines involves the solution of a system of equations with a
determinant of zero.  I suspect I will need to use singular value
decomposition.  Before I become embroiled in this task I am wondering
if anyone has tackled this problem before and would be willing to
share references and/or public domain routines for its solution.

Thanks.

Don Neth
neth@bme.ri.ccf.org
