This thesis compared gamut compression of color separations with
gamut compression in U*V*W* color space (the precursor of CIELUV). Non-parametric
tests were used to evaluate the results of subjective comparisons: which
approach was preferred and which was considered closest to the original.
The U*V*W* gamut-compression algorithm did a linear lightness compression,
followed by a chroma compression along lines of constant hue in U*V*W*
space. All chroma values were compressed linearly, as suggested by Evan's
Consistency Principle and Bartleson's Relativity Principle. The gamut compression
was image dependent and in the case of U*V*W* compression, modeled the
hardcopy output gamut used in the experiments as an irregular polyhedron.
Gamut compression in U*V*W* resulted in reproductions that were either
preferred or judged closer to the original. Lightness sharpening was added
in a post-processing step to increase the sharpness and the apparent contrast
and chroma of the gamut-compressed images.