F(x,y) = I(x,y)*R(x,y)
where I is the spatially slowly varying contents in an image (depending
on the illuminant) and R is a spatially rapidly varying function containing
the image detail itself. The I component is the one controlling the dynamic
range, which needs to be scaled to obtain a match in the lightness ranges
of the two gamuts (a technique for extracting the two components is given
in the paper). Chroma compression is carried out using a piecewise linear
function along lines of constant L* and hue angle. Finally any colours
which are still out of gamut after the first two stages are clipped onto
the gamut boundary along lines of constant lightness and hue angle.
The technique proposed here is similar to the technique proposed by
Johnson (1979) with the only difference being lightness compression. The
idea of gamut compression using image processing techniques is an interesting
one and could be further investigated to include chroma compression as
well.