Fabrice,
what I did over 25 years ago is to find the maximum degree of axial
symmetry of patterns by extensive correlation.
For pure 2D axial symmetry, the symmetry axis is defined by an angle and
a shift in _one_ dimension.
Therefore you have to determine a (generalized) correlation function
that depends on these two parameters. In other words, "try" pattern
reflections with respect to all possible symmetry axes, where "try"
means the computation of a similarity measure, e.g. correlation
coefficient. The similarity measure is the degree of axial symmetry for
the respective axis...
I'm not aware of any ImageJ-plugin for this task. However you may write
a macro or plugin that uses the ImageJ-plugin "TransformJ".
Please be aware that a reflection with respect to an arbitrary single
axis, followed by all rotations and shifts in x and y (3 parameters)
gives you the degree of _glide_ reflection symmetry.
HTH
Herbie
::::::::::::::::::::::::::::::::::::::::
On 27.12.14 11:21, Fabrice Senger wrote:
> Dear all,
>
> I'm struggling for a while with a particular question. Is there a tool
> or an approach to define the symmetry of a signal distribution ? What
> I'm interested in is to find the symmetry axis as a line.
>
> Thanks for your support,
>
> Fabrice.
>
> --
> ImageJ mailing list:
http://imagej.nih.gov/ij/list.html>
--
ImageJ mailing list:
http://imagej.nih.gov/ij/list.html
Locked
