Name: WlzHistogramFitPeaks - fits a series of Gaussian distributions to a histogram.

Synopsis

WlzHistogramFitPeaks [-h] [-y] [-o<out file>] [-s#] [-t#] [-l#]
                     [<in object>]

Options

-h	Help, prints usage message.
-l	Log-liklihood fitting tolerance.
-n	Number of Gaussian distributions.
-o	Output data file name.
-s	Minimum Gaussian sigma (standard deviation).
-t	Threshold value for peak height.
-y	Synthesize a new histogram from the mixture of Gaussians, by default just the statistics are output.

Description

Fits a mixture of Gaussian distributions to the given Woolz histogram using a maximum liklihood estimate. If given the number of Gaussian distributions is zero (default) then Gaussians will be fitted to peaks found using WlzHistogramFindPeaks. Either the statistics of the Gaussian mixture are output in the format

  <mu (mean)> <sigma (std dev)> <alpha (height)> <area>

one record for each fitted Gaussian, or a synthesized histogram built using the statistics is output. Objects/data are read from stdin and written to stdout unless the filenames are given.

Examples

WlzHistogramFitPeaks -s1.0 -t 10.0 myhist.wlz

The input Woolz histogram object is read from myhist.wlz. Histogram peaks are detected using a Gaussian smoothing filter with a sigma value of 1.0 and a threshold value of 10.0, the maximum liklihood mixture of Gaussians is computed and the mixture statistics are written to the standard output.

File: WlzHistogramFitPeaks.c

See Also: WlzIntro(1) WlzHistogramObj(1) WlzHistogramFitPeaks(3)