An RGB color management concept based on an improved gamut mapping algorithm

TLDR: This paper presents a practical implementation addressing both an eficient color transformation and a sophisticated gamut mapping of color reproduction on specific devices.

Abstract: In principal, color reproduction on specific devices can be divided into transformation of color coordinates and the adaptation of colors to a device, called gamut mapping. The well established ICC color management neglects the latter aspect. This paper presents a practical implementation addressing both an efficient color transformation and a sophisticated gamut mapping.

Figures

Table 1. Performance of building WCS grid

Table 4. Average errors with working color space CIELAB

Table 3. Performance of gamut mapping with default settings

Table 2. Performance of backtransform

Figure 5. Results of psycho-visual test

Frequently asked questions

Q1. What contributions have the authors mentioned in the paper "An rgb color management concept based on an improved gamut mapping algorithm" ?

This paper presents a practical implementation addressing both an efficient color transformation and a sophisticated gamut mapping. 

Q2. What future works have the authors mentioned in the paper "An rgb color management concept based on an improved gamut mapping algorithm" ?

In the future, their prototype will be improved in a number of ways. 

Q3. What is the main drawback of ICC-CMS?

The main drawback of ICC-CMS is that it is based on precomputed data structures (mostly lookup-tables), which means that the involved GM has to be precomputed, too. 

Q4. What is the main problem of the color management system?

Due to the limited range of colors reproducible by each device, an adaptation of the input colors, so-called gamut mapping (GM ),1 is usually involved.15 years ago, the International Color Consortium developed an open distributed software standard (the ICC color management system or ICC-CMS) to universally transform between color spaces, in particular device specific ones. 

Q5. How do the authors find the containing cube?

The containing cube is found by rounding the working color space coordinates of pt down to the next working color space grid vertex. 

Q6. How many images were mapped to the ISO gamut?

A set of a hundred images was mapped to the ISO newspaper gamut with each of the algorithms, resulting in a test set of 500 images. 

Q7. What is the way to map a color space?

Since ICC-CMS normally considers color spaces like sRGB or ECI RGB as input gamut, the ICC-CMS is always in a worst case situation. 

Q8. How many segments are split from pole to pole?

First the space, which the authors treat as a sphere, is partitioned into 808 segments along the equator and then each of these is further split into 404 segments from pole to pole. 

Q9. how many pixels can be mapped per second?

The backtransform is fast, it averages at about 1.13µs = 1.13 · 10−6s per pixel in the current implementation, which means about 890’000 pixels can be backtransformed per second (c.f. table 2). 

Q10. What is the definition of the term "Gamut mapping"?

SPIE-IS&T/ Vol. 7241 724118-8Downloaded from SPIE Digital Library on 25 Nov 2009 to 129.132.191.129. Terms of Use: http://spiedl.org/termsGamut mapping is clearly the bottleneck of the application: since it is filter-based, for every pixel being mapped also a certain amount of its neighborhood is taken into account. 

Q11. What are the representative representations of the gamut surface?

These representatives define the gamut surface, and are used for in-gamut queries as well as compression in gamut mapping (see section 4). 

Q12. How many pixels can be mapped per second?

Assuming the mapping was done with the default parameters described in section 4.3, it averages at about 11.14µs = 11.14 · 10−6s per pixel, meaning about 90’000 pixels can be mapped per second (c.f. table 3). 

Q13. What is the potential for improvement in ICC-CMS?

Avoiding the drawbacks of ICC-CMS as a distributed software standard, the authors were able to combine a sophisticated gamut mapping algorithm with a simple interpolation technique for the color transform. 

Q14. What is the way to improve color management?

SPIE-IS&T/ Vol. 7241 724118-10Downloaded from SPIE Digital Library on 25 Nov 2009 to 129.132.191.129. Terms of Use: http://spiedl.org/termsIn this paper the authors demonstrated that there are practical alternative color management systems with better visual performance.