Thresholding in 2D Color Space
Thresholding in multiple dimension can be explained by overlapping the hypercubes defined by thresholds with the distributions. In the following graph hypercubes are rectangles defined by selecting 80% of a distribution in each band.
These graphs lead to the concept of dimensionality evaluation for color classification. Dimensionality analysis is the answer to the question "How many color bands from the RGB signal are needed ?".