hi,
did anyone figure out how to find the mean orientation from a set of rotations ? It is not a simple mean operator on quaternions as q and - q represent the same orientation.
Keren.
You just have to find the choice of sign that brings them closest. Adding a function would be good.
On Jun 6, 2009, at 5:23 PM, Keren Lasker kerenl@salilab.org wrote:
> hi, > > did anyone figure out how to find the mean orientation from a set of > rotations ? It is not a simple mean operator on quaternions as q and > -q represent the same orientation. > > Keren. > _______________________________________________ > IMP-dev mailing list > IMP-dev@salilab.org > https://salilab.org/mailman/listinfo/imp-dev
(guessing from iphone ?? ;) ) i would add this function. what do you mean by bring them closest ? that I should prefer x+y over x-y if: dist(x,y) < dist(x,-y) ?
On Jun 6, 2009, at 5:58 PM, Daniel Russel wrote:
> You just have to find the choice of sign that brings them closest. > Adding a function would be good. > > > > On Jun 6, 2009, at 5:23 PM, Keren Lasker kerenl@salilab.org wrote: > >> hi, >> >> did anyone figure out how to find the mean orientation from a set >> of rotations ? It is not a simple mean operator on quaternions as q >> and -q represent the same orientation. >> >> Keren. >> _______________________________________________ >> IMP-dev mailing list >> IMP-dev@salilab.org >> https://salilab.org/mailman/listinfo/imp-dev > _______________________________________________ > IMP-dev mailing list > IMP-dev@salilab.org > https://salilab.org/mailman/listinfo/imp-dev
Yes. Basically, you are bringing things into the tangent space around one of the points. The projection onto this subspace would be the closest of -+x.
On Jun 6, 2009, at 6:04 PM, Keren Lasker kerenl@salilab.org wrote:
> (guessing from iphone ?? ;) ) > i would add this function. > what do you mean by bring them closest ? > that I should prefer x+y over x-y if: > dist(x,y) < dist(x,-y) ? > > On Jun 6, 2009, at 5:58 PM, Daniel Russel wrote: > >> You just have to find the choice of sign that brings them closest. >> Adding a function would be good. >> >> >> >> On Jun 6, 2009, at 5:23 PM, Keren Lasker kerenl@salilab.org wrote: >> >>> hi, >>> >>> did anyone figure out how to find the mean orientation from a set >>> of rotations ? It is not a simple mean operator on quaternions as >>> q and -q represent the same orientation. >>> >>> Keren. >>> _______________________________________________ >>> IMP-dev mailing list >>> IMP-dev@salilab.org >>> https://salilab.org/mailman/listinfo/imp-dev >> _______________________________________________ >> IMP-dev mailing list >> IMP-dev@salilab.org >> https://salilab.org/mailman/listinfo/imp-dev > > _______________________________________________ > IMP-dev mailing list > IMP-dev@salilab.org > https://salilab.org/mailman/listinfo/imp-dev
participants (2)
-
Daniel Russel -
Keren Lasker