* A bright structure on a dark backgroundДЄrray.getStatistics(y, ymin, ymax, ymean, ystdDev) * A line ROI drawn across a structure of interest for the Breit-Wigner this is just Gamma (by definition) and for the Gaussian: exp(-x2/2sigma2) 1/2 x2/2sigma2. * Olivier Burri, Romain Guiet, EPFL-SV-BIOP * Example finding length of line-like structures using supergaussian fitting and FWHM Hopefully this can be useful as a sarting point as well as the other approaches suggested. IвЂ™ve attached a macro that takes a line profile and tries to fit a supergaussian through the data and provides the FWHM. Using it in the curve fitter gives a nice fit of this artificial data You can find the formula there or on Wikipedia ДЄnd the calculation of the FWHM is rather simple to derive based on the one from the regular exponential, Our first dealing with the supergaussian with my colleague stemmed from a project a couple of years ago that led us to this paper and led us to try it out on ImageJвЂ™s Curve Fitter. In the case of these line-like objects and distributions, a good candidate fit is the supergaussian. If you want to draw these manually and then assess their distance, approach is on the right track, except that the FWHM can be defined for functions other than the plain gaussian. With smooth constant lines like the one showed, this can be useful, but might redirect the problem to вЂњHow do we get to draw a line profile across the object in the first place?вЂќ Hi the FWHM can be applied as long as you consider that drawing a line trough the half maximum of a profile and measuring its length (or infering it from calculation) is meaningful.

Full width at half max