Methods and algorithms for finding central symmetry in binary images
Jaccard index for estimating the degree of symmetry
Real world objects that are perceived as symmetrical may not be so in the strict sense.
The reasons are natural defects of the figure, optical distortions, the presence of noise and others.
A suitable way to estimate the measure of symmetry of the figure (set of pixels), which
returns a value in the range [0,1], is the Jaccard index:
A is the original shape, A' is the rotated shape.
Obviously, rotation relative to different points provides different values of the Jaccard
index. Therefore, the task is to find a focus that maximizes the Jaccard index when
turning at given angles.