Something about Voronoi.. (maybe a feature idea though)

[Rainer M. Engel]

I wonder if something like this is possible with ImageJ.
http://dl.dropbox.com/u/414433/imagej/voronoi.jpg

The blur in the last step is a uniform gauss and not an interpolation by
distance of the existing points. I only wanted to show the idea.

With this method I would like to get a smooth gradient between points
and their corresponding measurement value (i.e. color).
Maybe Georgi Woronoi would have some colourful fun with it either, I
don't know...

May be a feature idea. Any further ideas/opinions?

Regards,
Rainer


--
Rainer M. Engel, Dipl. Digital Artist
scientific|Media GbR
Pichelsdorferstr. 143
D-13595 Berlin

[Stephan Saalfeld]

There is an ImgLib2 test that does exactly that.  Check it out and try:

http://fiji.sc/cgi-bin/gitweb.cgi?p=imglib.git;a=blob;f=imglib2/ij/src/test/java/net/imglib2/nearestneighbor/KNearestNeighborSearchBehavior.java

rendering a picture of myself sampled at 1000 random locations using
several different weights and neighborhoods for a distance weighted
interpolation.  Check the method drawWeightedByDistance at line 79.

ImgLib2 is in Fiji.  Alternatively, you can extract the imglib2 jars
from the Fiji repositories and use them in virgin ImageJ.

Best,
Stephan


On Tue, 2012-01-24 at 21:45 +0100, Rainer M. Engel wrote:

> I wonder if something like this is possible with ImageJ.
> http://dl.dropbox.com/u/414433/imagej/voronoi.jpg
>
> The blur in the last step is a uniform gauss and not an interpolation by
> distance of the existing points. I only wanted to show the idea.
>
> With this method I would like to get a smooth gradient between points
> and their corresponding measurement value (i.e. color).
> Maybe Georgi Woronoi would have some colourful fun with it either, I
> don't know...
>
> May be a feature idea. Any further ideas/opinions?
>
> Regards,
> Rainer
>
>

[Rainer M. Engel]

Thank you Stephan for this tip..

fiji is fine and I'll have a look at that.



Am 30.01.2012 10:15, schrieb Stephan Saalfeld:

> There is an ImgLib2 test that does exactly that.  Check it out and try:
>
> http://fiji.sc/cgi-bin/gitweb.cgi?p=imglib.git;a=blob;f=imglib2/ij/src/test/java/net/imglib2/nearestneighbor/KNearestNeighborSearchBehavior.java
>
> rendering a picture of myself sampled at 1000 random locations using
> several different weights and neighborhoods for a distance weighted
> interpolation.  Check the method drawWeightedByDistance at line 79.
>
> ImgLib2 is in Fiji.  Alternatively, you can extract the imglib2 jars
> from the Fiji repositories and use them in virgin ImageJ.
>
> Best,
> Stephan
>
>
> On Tue, 2012-01-24 at 21:45 +0100, Rainer M. Engel wrote:
>> I wonder if something like this is possible with ImageJ.
>> http://dl.dropbox.com/u/414433/imagej/voronoi.jpg
>>
>> The blur in the last step is a uniform gauss and not an interpolation by
>> distance of the existing points. I only wanted to show the idea.
>>
>> With this method I would like to get a smooth gradient between points
>> and their corresponding measurement value (i.e. color).
>> Maybe Georgi Woronoi would have some colourful fun with it either, I
>> don't know...
>>
>> May be a feature idea. Any further ideas/opinions?
>>
>> Regards,
>> Rainer
>>