Re: Accurate Gaussian Blur

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view
|

Re: Accurate Gaussian Blur

Michael Schmid
Hi,

the ImageJ built-in Gaussian Blur as well as the 'Accurate Gaussian Blur' plugin do the very basic operation described in
  http://en.wikipedia.org/wiki/Gaussian_blur

You will find the equation there.  The main difference between the two is using double precision arithmetics in the 'Accurate Gaussian Blur', which eliminates some very small rounding errors (about 10^-7 of the pixel value).  Usually, the difference between the built-in and the 'accurate' Gaussian blur are not noticeable; but the 'accurate' one is slightly slower.

The inner workings are a bit more complicated than described in the Wikipedia page, but that should not bother you - the result is the same.  If you are really interested in these details, see the documentation in the source code and this thread in the mainling list:
  https://list.nih.gov/cgi-bin/wa.exe?A2=ind1403&L=IMAGEJ&P=R35207&1=IMAGEJ

By the way, please send such questions to the mailing list (I also copy the response there).  I am not the ImageJ help desk ;-)

Michael
________________________________________________________________
On Jun 12, 2014, at 07:22, rramakrishnan wrote:

> Dear Professor Michael Schmid,
>
> I am Ramakrishnan R, a PhD student at Kyung Hee University in South Korea.
> I used the Image J plugin "Accurate Gaussian blur" for my experiments in microscopy. It would be of my help if I can get the Gaussian equation that was used in the plugin. Kindly provide me with the equation of the gaussian function used in the plugin and also your reference paper where the equation and the plugin are postulated.
>
> Thank You,
> Yours Truly,
> Ramakrishnan R,
> Advanced Convergence Analytical chemistry Lab (ACACL),
> Kyung Hee University, Republic of Korea.
> Lab Link : http://nbacl.khu.ac.kr/
>

--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html