Quantifying Colour Appearance. Part II. Testing Colour Models Performance Using Lutchi Colour Appearance Data

M. Ronnier Luo,A. A. Clarke,P. A. Rhodes,A. Schappo,S. A. R. Scrivener,C. J. Tait
DOI: https://doi.org/10.1002/col.20255
IF: 1.668
1991-01-01
Color Research & Application
Abstract:AbstractThe acquisition of the LUTCHI Colour Appearance Data has been described in Part I of this article. Having obtained the data, they were used to test the accuracy of prediction for various colour spaces and models. The results clearly indicate that Hunt's 91 model gives the best fit to the visual results of all the models studied. Hunt's 91 has been further refined to improve the fit to the colourfulness results, and this refined model has been designated Hunt‐ACAM (ACAM being the Alvey Colour Appearance Model). The error of prediction from Hunt‐ACAM is close to the typical error that is seen to occur between individuals' results and the mean visual results. This performance is considered to be very satisfactory, and the model is therefore believed to provide a reasonably accurate way of evaluating colour fidelity for various colour reproduction systems. Various chromatic‐adaptation transformations were also compared with three sets of corresponding chromaticities derived from the results of experiments conducted under four conditions of chromatic adaptation. The results are in reasonable agreement with those obtained by Helson et al. [Illum. Eng. 47, 221–233 (1952)] and Lam and Rigg [Ph.D. thesis, University of Bradford (1985)]. All results indicate that the Bradford and Hunt‐ACAM transformations perform the best and the second best, respectively, of all the selected formulae. The current CIE recommendation does not perform as well as expected.
What problem does this paper attempt to address?