1. Introduction
In the research of computer vision, accurate transmission and reproduction of color information are the premise of the
analysis and reproduction of natural sceneries. Since the recording and representing abilities of various media are very
different (mainly because different media have different color gamut), color information is often lost or misrepresented in
the process of image transmission and reproduction [1].
The function of the Color Management System (CMS) is to guarantee the accurate color transmission and reproduction
among different media [10]. The basic mechanism of the current CMS is based on CIE LAB Standard Color Model as shown in
Fig. 1. The color information of input media is first transformed into the Standard Color Model CIE LAB [1]. Then according
to recording and representing ability of the output media, color gamut mapping is implemented from the input media to
the output media in CIE LAB model. Finally, the color information is transformed from CIE LAB model to the output media.
The main problem of using CIE LAB as a Standard Color Model to carry out gamut mapping is: although the CIE LAB
color model can render all colors, differences among colors do not conform to human vision. Human eyes’ feeling of the
color differences is usually denoted by the space distance in the color model. In CIE LAB model, the space distance between
colors is different from the perception of human eyes. Generally, the color model which does not conform to human vision
is called uneven color model, and it is quite dangerous to carry out gamut mapping in an uneven color model. If it is used
to map the color gamut from a big gamut media to a small gamut media, it will not only cause the uneven compression of
the gamut, but also cause the color in the sensitive area of human eyesight to wrongly undergo a large-scale compression.
Therefore, choosing a proper standard color model as an intermediate color gamut mapping platform is the key technology
for correct color information transmission and reproduction among various media [6]. This paper proposes to use iCAM
color appearance model as the standard model in CMS and to create a CMS based on iCAM color appearance model. The
experiment result has proven that the CMS based on iCAM appearance model is superior to the current CMS based on CIE
LAB model.
✩ The research is supported by Nature Science Foundation of Jiangsu Province BK2006210 C