#219 Color Clustering and Segmentation

[Katherine Ann Scott]

Use kmeans or a similar algorithm to generate a statistical summary of the colors in an image. Optionally create a feature type that describes this information for tracking and display. The kmeans analysis would figure out the k most prominent colors in an image an return the values as a palette feature. Make it possible to "reduce" an image to its constituent palette features, or segment all of the objects that have/don't have that color.

Example
p = Palette()
p = img.getPalette(k=10,getDistribution = true)
blobs = img.segmentFromPalette(p[0],p,thresh=10) # get blobs with palette color zero.
img.sideBySide(p.toImage(doDist=True,size=(30,200),type="vertical"))# do a side by side of the palette drawn proportional to the distribution of colors