Re: 2D gaussian fit of puncta

[dpoburko]

Hi All,

   Has anyone has any success in implementing a 2-D Gaussian fit in
ImageJ? I am looking for a method of localizing the center and FWHM of
fluorescent puncta by implementing a simultaneous 2-D fit. Some
colleagues have suggested that this would be trivial in MatLab, but it
seems like this would be a valuable tool to be able to run in ImageJ
(independent of other programs). I am currently using Adrian's FWHM.
It's quite good, but I would like to go beyond the independent X and Y
fits. If anyone has this implemented or has run into pitfalls in past
attempts, I'd love to hear about it.

Cheers,
Damon

--
Post Doctoral Fellow
Tsien Lab, Molecular&  Cellular Physiology
Stanford University

[Nico Stuurman-4]

Hi Damon,

> Has anyone has any success in implementing a 2-D Gaussian fit in ImageJ? I am looking for a method of localizing the center and FWHM of fluorescent puncta by implementing a simultaneous 2-D fit. Some colleagues have suggested that this would be trivial in MatLab, but it seems like this would be a valuable tool to be able to run in ImageJ (independent of other programs). I am currently using Adrian's FWHM. It's quite good, but I would like to go beyond the independent X and Y fits. If anyone has this implemented or has run into pitfalls in past attempts, I'd love to hear about it.

With much help of Arthur Edelstein, I wrote a Gaussian Fit plugin for ImageJ.  It is still a work in progress (hence, no previous announcement), but seems to work quite well and is orders of magnitude faster than Gaussian Fit in Matlab.  You can find the code and rudimentary instructions at:

http://valelab.ucsf.edu/~nico/ijplugins/GaussianFit.html

Note that the Matlab import script is currently broken, will be fixed soon.

Best,

Nico