Bidirectional individual corresponding colors data. Issue 5 (1st April 2018)

• Record Type: Journal Article
• Authors: Cai, Shengyan; Fairchild, Mark D.
• Is Part Of: Color research & application. Volume 43:Issue 5(2018:Oct.)
• Page Start: 643
• Page End: 654
• Publication Date: 2018-04-01
• Subjects: accuracy -- bidirectional adaptation -- corresponding colors data -- individual differences -- precision; Color -- Periodicals; 535
• DOI: 10.1002/col.22228
• Languages: English
• ISSNs: 0361-2317

Abstract

Abstract: The precision of corresponding colors data is critical to improve chromatic adaptation models. This research aims to obtain high precision and individually resolved corresponding colors data, as well as to look into the reversibility of chromatic adaptation. The experiment included 2 phases. The illumination changes were simply reverses of one another. The mean results of the forward phase across 9 observers were adopted as the test colors in the reverse phase. The CIEDE2000‐unit standard deviations of the results across 9 observers in the forward and reverse phases are 2.81 and 2.56, respectively. The standard error of the mean of the results of individual observers and across 9 observers of the 2 phases are all less than 1.0 CIEDE2000 unit, very significantly more precise than previously reported data. Nevertheless, by examination of the bidirectional procedures, it is concluded that there might be need to improve accuracy or contemplate a reality in which chromatic adaptation is not a reversible process (ie, has a significant hysteresis effect). The mean biases between the test colors of the forward phase and the visual results of the reverse phase for each individual across 5 colors are averaged at 7.2 CIEDE2000 units. The average value for overall data is nearly 6.0 CIEDE2000 units. This indicates that chromatic adaptation in this experiment is not reversible. The discrepancies are rather large and need to be explored further. One possible, but by no means definitive, explanation is that the biases are might be caused by memory distortions in the experiment.