Comparison and Summary Statistics¶
Tools for calculating derived quantities (image2 images or reductions) based on an image. Should all be in the form def q(image) to be accessible as image.
- imtools.stats.dssim(var1, var2)¶
Image dissimilarity DSSIM is 1/abs(SSIM) - 1
- imtools.stats.dssims(image1, image2)¶
DSSIM for each variable
- imtools.stats.mle(var1, var2)¶
Mean “Linear” error (normalized L1 norm) Follows GRRT paper definition in dividing by the sum(var) of first image
- imtools.stats.mles(image1, image2)¶
MSE of this image vs animage2. Follows GRRT paper definition in dividing by the sum(var) of first image
- imtools.stats.mse(var1, var2)¶
MSE between images. Follows GRRT paper definition in dividing by the sum(var) of first image
- imtools.stats.mses(image1, image2)¶
MSE for each variable.
- imtools.stats.ncc(var1, var2)¶
Normalized Cross-Correlation NCC (like ZNCC without mean subtraction)
- imtools.stats.nccs(image1, image2)¶
NCC for each variable
- imtools.stats.polar_abs_integrated(image1, image2)¶
Absolute differences in the total flux, integrated & averaged linear polarization fractions, and circular polarization fraction.
- imtools.stats.polar_rels_integrated(image1, image2)¶
Relative differences in the total flux, integrated & averaged linear polarization fractions, and circular polarization fraction. (That is, a “polar” breakdown with magnitudes instead of a “Cartesian” breakdown with Stokes Parameters)
- imtools.stats.rel_integrated(var1, var2)¶
Relative difference of the sum of a variable
- imtools.stats.rels_integrated(image1, image2)¶
Integrated relative errors for each variable
- imtools.stats.ssim(var1, var2)¶
Image structural similarity SSIM as defined in Gold et. al, eq. 14
- imtools.stats.ssims(image1, image2)¶
SSIM for each variable
- imtools.stats.zncc(var1, var2)¶
Zero-Normalized Cross-Correlation ZNCC (normalized correlation of deviation from mean)
- imtools.stats.znccs(image1, image2)¶
ZNCC for each variable