Re: [Toon-members] tag tag/absorient.h src/absorient.cpp test/sim.cpp

[Gerhard Reitmayr]

Hi Pekka,

Thanks for the reference, I didn't know that paper. I will probably implement 
it, as it will provide better guarantees. I have not seen a reflexion in real 
life, but I can imagine that it may happen.

cheers,
  Gerhard

On 27 Jun 2011, at 07:26, Pekka Paalanen wrote:

> On Sat, 25 Jun 2011 09:12:21 +0000
> Gerhard Reitmayr <address@hidden> wrote:
> 
>> CVSROOT:     /cvsroot/toon
>> Module name: tag
>> Changes by:  Gerhard Reitmayr <gerhard>      11/06/25 09:12:21
>> 
>> Modified files:
>>      tag            : absorient.h 
>>      src            : absorient.cpp 
>> Added files:
>>      test           : sim.cpp 
>> 
>> Log message:
>>      added estimating a similarity transform (se3 + scale) between
>> two 3D point sets after Horn
> 
> Hi,
> 
> I see you refer to Horn. Are you aware of the paper by Shinji Umeyama,
> Least-Squares Estimation of Transformation Parameters Between Two Point
> Patterns, IEEE TPAMI 13(4), April 1991?
> 
> I did not check your code nor which Horn's paper you refer to, but
> Umeyama claims to avoid the problem of reflections in comparison to
> Horn. IIRC the biggest difference is that Umeyama uses SVD instead of
> eigenvalue decomposition.
> 
> Just in case you might find it interesting.
> 
> 
> Cheers.
> 
> -- 
> Pekka Paalanen, M.Sc. (Eng.)   LUT/IT, room 6505
> doctoral student               http://www.it.lut.fi/project/rtmosaic/
> address@hidden          http://www.iki.fi/pq/

--
Gerhard Reitmayr
Institute for Computer Graphics and Vision
http://www.icg.tugraz.at/Members/gerhard
tel: ++43 316 873 5082