CIE Division 8 - TC8-03: Survey of Gamut Mapping Papers

Morovic & Luo (1997a)

This paper describes the evaluation of the following seven algorithms, most of which were published before:

1. LLIN - uniform linear lightness compression followed by linear compression of chroma along lines of constant lightness (Johnson, 1979).
2. LNLIN - same as LLIN with the difference that chroma was compressed non-linearly using a soft-clipping function (Stone and Wallace, 1991).
3. LCLIP - same as LLIN with the difference that chroma was clipped (Sara, 1984).
4. LSLIN - uniform linear lightness compression followed by compression towards the centre of the lightness axis (Laihanen, 1987).
5. JOHNSON - a misinterpreted version of the CARISMA algorithm (CARISMA, 1992).
6. CUSP - compression towards the lightness of the cusp on the lightness axis.
7. SLIN - compression towards the centre of the lightness axis (note, that this was also implemented in LLAB (Luo and Morovic, 1996)).

Note, that the last two of the above algorithms did not have an overall lightness compression stage, so as to verify whether its use is always necessary. Printed reproductions were compared to original CRT images using the pair comparison technique where observers were asked to judge the accuracy of reproduction. The experimental setup was such that the luminances and chromaticities of the two media were approximately equal. Overall, the results of this experiment suggested that the CUSP, SLIN and SLINLLAB algorithms perform significantly better than the other GMAs. However, the experimental results for individual colour areas in the used test images also showed that these algorithms fail for the neutral axis and for dark colours. It is also of interest to note that the LCLIP method, which in other studies performed well, did particularly badly in this experiment. This is probably due to the gamut difference being larger here since CRT images were mapped to an inkjet printer having a lightness range which was approximately 20 per cent smaller.

Last updated: 17 August 1999 by Jan Morovic