Corresponding-Colour Datasets - Luo and Rhodes (1999)

Source URL: https://web.archive.org/web/20031123133629/http://colour.derby.ac.uk:80/colour/info/catweb/
Source DOI: https://doi.org/10.1002/(SICI)1520-6378(199908)24:4%3C295::AID-COL10%3E3.0.CO;2-K

M. Ronnier Luo and Peter A. Rhodes

Colour & Imaging Institute
University of Derby
Derby
England

INTRODUCTION

A chromatic adaptation transform is capable of predicting corresponding colours. Corresponding colours are described by two sets of tristimulus values that give rise to the same perceived colour when the two samples are viewed under test and reference light sources or illuminants. The two light sources or illuminants differ in terms of their colour temperatures (or chromaticity coodinates). A chromatic adaptation transform can be effectively used for numerous industrial applications such as the evaluation of colour inconstancy for surface samples, the calculation of colour difference between pairs of samples assessed under non-daylight sources or illuminants, the provision of a colour rendering index for assessing the quality of light sources, or the prediction of coloured images across different sources or illuminants.

In October 1998, the CIE formed a new technical committee, TC 1-52, on Chromatic Adaptation Transforms during its interim meeting in Baltimore, USA with Professor M. R. Luo as its chairman. The objective of this committee is to review certain chromatic adaptation transforms with a view to making a CIE recommendation. The performance of chromatic adaptation transforms is normally evaluated using corresponding-colour experimental data sets in which each colour is defined by two sets of tristimulus values under two illuminants. Many experiments were carried out using a variety of psychophysical methods under different viewing conditions. A comprehensive collection of these data sets has been accumulated by Luo and Hunt [1] for the purposes of deriving and evaluating the CIE colour appearance model, CIECAM97s [2], and the CMC chromatic adaptation transform, CMCCAT97 [3]. The Committee has decided to make these data sets available via the Internet for public assessment. This task has been completed and the resulting database is now available via the world wide web at http://colour.derby.ac.uk. Researchers or industrialists are welcome to acquire this database for further study. This paper gives a brief description of each data set and describes the format of the data.

EXPERIMENTAL DATA SETS

Fourteen data sets have been accumulated from nine sources [4-11]: the Color Science Association of Japan (CSAJ), Helson, Lam and Rigg, LUTCHI, Kuo and Luo, Breneman, Braun and Fairchild, and McCann. Each data set includes a number of corresponding-colour pairs in which both colours in a pair appear the same when each is viewed under different viewing conditions. Table I summarises the experimental conditions in each data set including the number of phases (as defined by a set of viewing conditions), the number of corresponding-colour pairs and the viewing parameters used. The parameters considered are the light sources used for the test and reference conditions, illuminance (lux), the luminance factor of the neutral background (Y%), sample size, media and psychophysical method.

The CSAJ [4] data was divided into three sets: -C, -Hunt and -Stevens according to studies on chromatic adaptation, Hunt and Stevens effects respectively. The Helson [5], Lam and Rigg [6] data sets include corresponding colours between the test source (A) and reference source (D65). The LUTCHI [7] data includes three sets - A, D50 and WF - which are the test illuminants against a reference D65 simulator. Similarly, there are two sets for Kuo and Luo [8] data: A and TL84, which are the test light sources against a reference D65 simulator. The only data set based upon transparent media in this category is the Breneman [9] data which was divided into two sets: -C and -L according to investigations on chromatic adaptation and illuminance effects respectively. The Braun and Fairchild [10] data was accumulated by asking observers to adjust monitor colours to match those presented on reflection prints. The McCann [11] data were obtained by investigating the chromatic adaptation effect using a Mondrain figure viewed under highly chromatic test illuminants with low illuminances. Its original data was further analysed to obtain corresponding tristimulus values by Nayatani et al [12].

In total, 746 corresponding-colour pairs were gathered from experiments involving 38 phases of viewing conditions. The psychophysical methods used are haploscopic matching, memory matching and magnitude estimation.

Table I: Summary of the corresponding-colour data sets

| Data Set | No. of Phases | No. of Samples | Illuminant | | Illuminance(lux) | Background(Y%) | Sample Size | Medium | Experimental Method |
|-------------------|---------------|----------------|------------|---------------|------------------|----------------|-------------|-------------|---------------------|
| | | | Test | Ref. | | | | | |
| CSAJ-C | 1 | 87 | D65 | A | 1000 | 20 | S | Refl. | Haploscopic |
| CSAJ-Hunt | 4 | 20 | D65 | D65 | 10-3000 | 20 | S | Refl. | Haploscopic |
| CSAJ-Stevens | 4 | 19 | D65 | D65 | 10-3000 | 20 | S | Refl. | Haploscopic |
| Helson | 1 | 59 | D65 | A | 1000 | 20 | S | Refl. | Memory |
| Lam & Rigg | 1 | 58 | D65 | A | 1000 | 20 | L | Refl. | Memory |
| Lutchi (A) | 1 | 43 | D65 | A | 1000 | 20 | S | Refl. | Magnitude |
| Lutchi (D50) | 1 | 44 | D65 | D50 | 1000 | 20 | S | Refl. | Magnitude |
| Lutchi (WF) | 1 | 41 | D65 | WF | 1000 | 20 | S | Refl. | Magnitude |
| Kuo & Luo (A) | 1 | 40 | D65 | A | 1000 | 20 | L | Refl. | Magnitude |
| Kuo & Luo (TL84) | 1 | 41 | D65 | TL84 | 1000 | 20 | S | Refl. | Magnitude |
| Breneman-C | 9 | 107 | D65, D55 | A, P, G | 50-3870 | 30 | S | Trans. | Magnitude |
| Breneman-L | 3 | 36 | D55 | D55 | 50-3870 | 30 | S | Trans. | Haploscopic |
| Braun & Fairchild | 4 | 66 | D65 | D30, D65, D95 | 129 | 20 | S | Mon., Refl. | Matching |
| McCann | 5 | 85 | D65 | R, Y, G, B | 14-40 | 30 | S | Refl. | Haploscopic |

DATA FILE DESCRIPTION

Table II summarises the data file names and number of samples in each experimental data set. For each file, the data were arranged in a fixed format. In the top row, there are six figures corresponding to the tristimulus values of the reference and test illuminants, i.e. Xr, Yr, Zr and Xt, Yt, Zt. There is only one figure in the second row denoting the number of samples in the file. The other rows are also arranged in the same manner as the first row. These are the corresponding tristimulus values under the reference and test illuminants for each sample.

Table II: The data file names and number of samples in each experimental data set

| Data Set | No. of Specimen | File Names |
|-------------------|-----------------|----------------|
| CSAJ-C | 87 | CSAJ.da.dat |
| CSAJ-Stevens | 5 | Steve.10.dat |
| | 5 | Steve.50.dat |
| | 5 | Steve.1000.dat |
| | 4 | Steve.3000.dat |
| CSAJ-Hunt | 5 | CSAJ.10.dat |
| | 5 | CSAJ.50.dat |
| | 5 | CSAJ.1000.dat |
| | 5 | CSAJ.3000.dat |
| Helson | 59 | helson.ca.dat |
| Lam & Rigg | 58 | lam.da.dat |
| Lutchi (A) | 43 | lutchi.da.dat |
| Lutchi (D50) | 44 | lutchi.dd.dat |
| Lutchi (WF) | 41 | lutchi.dw.dat |
| Kuo & Luo (A) | 40 | Kuo.da.dat |
| Kuo & Luo (TL84) | 41 | Kuo.dt.dat |
| Breneman-C | 12 | Brene.p1.dat |
| | 12 | Brene.p2.dat |
| | 12 | Brene.p3.dat |
| | 12 | Brene.p4.dat |
| | 11 | Brene.p6.dat |
| | 12 | Brene.p8.dat |
| | 12 | Brene.p9.dat |
| | 12 | Brene.p11.dat |
| | 12 | Brene.p12.dat |
| Breneman-L | 12 | Brene.p5.dat |
| | 12 | Brene.p7.dat |
| | 12 | Brene.p10.dat |
| Braun & Fairchild | 17 | RIT.1.dat |
| | 16 | RIT.2.dat |
| | 17 | RIT.3.dat |
| | 16 | RIT.4.dat |
| McCann | 17 | mcan.b.dat |
| | 17 | mcan.g.dat |
| | 17 | mcan.grey.dat |
| | 17 | mcan.r.dat |
| | 17 | mcan.y.dat |

SUMMARY

A data library has been established to include fourteen corresponding-colour data sets and is available at from http://colour.derby.ac.uk. It is publicly available for all interested parties who are doing research in the areas of colour appearance and chromatic adaptation. Users are encouraged to report new findings to the Chairman of CIE TC 1-52.

REFERENCES

LUO M.R. and HUNT R.W.G, Testing colour appearance models using corresponding-colour and magnitude-estimation data sets, Color Res. Appl. 23 147-153 (1998).

LUO M.R. and HUNT R.W.G, The structure of the CIE 1997 colour appearance model (CIECAM97), Color Res. Appl. 23 138-146 (1998).

LUO M.R. and HUNT R.W.G, A chromatic adaptation transform and a colour inconstancy index, Color Res. Appl. 23 154-158 (1998).

MORI L., SOBAGAKI H., KOMATSUBARA H. and IKEDA K., Field trials on the CIE chromatic adaptation formula, Proceedings of the CIE 22nd Session, 55-58 (1991).

HELSON H., JUDD D. B., and WARREN M. H., Object-color changes from daylight to incandescent filament illumination, Illum. Eng. 47, 221-233 (1952).

LAM K. M., Metamerism and colour constancy, Ph.D. Thesis, University of Bradford, 1985.

LUO M. R., CLARKE A. A., RHODES P. A., SCRIVENER, S. A. R., SCHAPPO A. and TAIT C.J., Quantifying Colour Appearance. Part I. LUTCHI Colour Appearance Data, Color Res. Appl., 16 166-180 (1991).

KUO W.G., LUO M.R., BEZ H. E., Various chromatic-adaptation transforms tested using new colour appearance data in textiles, Color Res. Appl., 21 313-327 (1995).

BRENEMAN E. J., Corresponding chromaticities for different states of adaptation to complex visual fields, J. Opt. Soc. Am. 4:6 1115-1129 (1987).

MCCANN J.J., MCKEE S. P. and TAYLOR T. H., Quantitative studies in retinex theory, Vision Res. 16 445-458 (1976).

BRAUN K.M. and FAIRCHILD M.D., Psychophysical generation of matching images for cross-media color reproduction, in IS&T and SID's 4th Color Imaging Conference: Color Science, Systems and Applications, 214-220, IS&T, Springfield, Va., (1996).

NAYATANI Y., TAKAHAMA K. and SOBAGAKI H., Prediction of color appearance of object colors in a complex visual field, J. Light & Vis. Env. 19 5-14 (1995).