基本的概念就是使用cluster SOP来群集色彩。通常cluster SOP就是说在位置矢量上执行K-Means群集。或者找出点的组，并且群集它们。不过色彩也是一个矢量。因此cluster SOP可以很容易的用于找到色彩组，并且将它们成组。如果你希望将一幅图像上的色彩减少为最显著的色彩，使用它就很方便了。

The basic Idea is to abuse the cluster SOP to cluster colors. Usually the cluster SOP is meant to perform K-Means Clustering on position vectors. Or: to find groups of points and cluster them. However a color also is a vector. So the cluster SOP can easily be used to also find groups of colors and cluster them. Really handy if you want to reduce the color palette of an image down to its most prominent colors.