StatMoments_Distance¶

class turbustat.statistics.StatMoments_Distance(image1, image2, radius=5, weights1=None, weights2=None, nbins=None, periodic1=False, periodic2=False, fiducial_model=None)[source] [edit on github]¶

Bases: object

Compute the distance between two images based on their moments. The distance is calculated for the skewness and kurtosis. The distance values for each for computed using the Hellinger Distance (default), or the Kullback-Leidler Divergence.

Unlike the other distance classes in TurbuStat, the computation of the histograms needed for the distance metric has been split into its own method. However, the change is fairly transparent, since it is called within distance_metric.

Parameters:

image1 : numpy.ndarray or astropy.io.fits.PrimaryHDU or spectral_cube.LowerDimensionalObject 2D Image. image2 : numpy.ndarray or astropy.io.fits.PrimaryHDU or spectral_cube.LowerDimensionalObject 2D Image. radius : int, optional Radius of circle to use when computing moments. This is the pixel size applied to the coarsest grid (if the datasets are not on a common grid). The radius for the finer grid is adjusted so the angular scales match. weights1 : numpy.ndarray or astropy.io.fits.PrimaryHDU or spectral_cube.LowerDimensionalObject 2D array of weights. Uniform weights are used if none are given. weights2 : numpy.ndarray or astropy.io.fits.PrimaryHDU or spectral_cube.LowerDimensionalObject 2D array of weights. Uniform weights are used if none are given. nbins : int, optional Number of bins to use when constructing histograms. periodic1 : bool, optional If image1 is periodic in the spatial boundaries, set to True. periodic2 : bool, optional If image2 is periodic in the spatial boundaries, set to True. fiducial_model : StatMoments Computed StatMoments object. use to avoid recomputing.

Methods Summary

create_common_histograms([nbins])	Calculate the histograms using a common set of bins.
distance_metric([metric, verbose, nbins, ...])	Compute the distance.

Methods Documentation

create_common_histograms(nbins=None)[source] [edit on github]¶

Calculate the histograms using a common set of bins. Only histograms of the kurtosis and skewness are calculated, since only they are used in the distance metric.

Parameters:	nbins : int, optional Bins to use in the histogram calculation.

distance_metric(metric='Hellinger', verbose=False, nbins=None, label1=None, label2=None, save_name=None)[source] [edit on github]¶

Compute the distance.

Parameters:	metric : ‘Hellinger’ (default) or “KL Divergence”, optional Set the metric to use compare the histograms. verbose : bool, optional Enables plotting. nbins : int, optional Bins to use in the histogram calculation. label1 : str, optional Object or region name for image1 label2 : str, optional Object or region name for image2 save_name : str,optional Save the figure when a file name is given.