The Design of Advanced Gamut Mapping Algorithms in Color Management Systems

Details:

Year: 2003
Pages: 18

Summary:

Currently the design for Color Management System (CMS) is based on the concept of Device Independent Color to implement the optimal color imaging across different imaging devices. However, the color appearance may not match between source and destination images although compared under the same viewing conditions. It is due to the dissimilar color gamuts existed between these two imaging devices used. Therefore it is needed to apply an optimal Gamut Mapping Algorithm (GMA) in a CMS process. The objective of this study is to derive GMAs to overcome the color-imaging problem caused by the mismatch of color gamuts in cross-media environments. Two imaging devices of a Hi-Fi monitor and an Inkjet printer were tested. The GMAs derived here, by taking into account of false boundary problem, directly carried out in a uniform 3D color space without any constraint, were categorized into two types of single-focal-point (SFP), and multiple-focal-points (MFP). Separately each type also carried out three approaches of clipping, linear compression and nonlinear compression. The SFP models used L*= 50 as the single focal point, whereas the MFP ones applied focal points repectively on three parts of highlight, mid-tone, and shadow. Before applying MFP method, a S-type scale method was implemented to change gamut distributions of originals tested. The S-type scale method let both the original and the reproduction gamuts have the same lightness range but optimally kept the original shadow areas of gradation. An advanced evaluation of models performance, by adopting a forced-choice paired-comparison experiment, was carried out using both of complex and computer-graphic color images. Results showed that the GMA, coupled with both of non-linear 3-D gamut compression and multiple-mapping-directions approaches, introduced better renditions than other five models tested. Especially it was more evident to complex images, but less evident to CG images having fewer gradations. Also the non-linear gamma-compression method minimized the loss of details occurred in clipping, while retained the merit of accurately reproducing most of common gamut colors.