Hi,
I had a discussion in sci.math about the derivation of finding a least
squares similarity transformation (uniform scaling + rotation +
translation also called a conformal affine transformation) between two
point sets. I actually was talking about the 2d problem, but got as a
response a dimension independent solution. I have only seen the 2d
derivation in papers and was excited, so I think this might of interest
to some of you too. As written and solved by Rob Johnson:
http://www.whim.org/nebula/math/conformalregress.html
There is also a similar derivation for the least squares affine
transformation.
http://www.whim.org/nebula/math/affineregress.html
(Although I'm still struggling with the step [1] -> [2] myself:))
--
http://kaba.hilvi.org
