How do Major Color-Difference Formulae Perform in the High Chroma Blue Region?

摘要

The objectives of this work were to develop a comprehensive visual dataset, NCSU-B2, around the CIE high chroma blue color centre and to use the new dataset as well as the low chroma blue dataset, NCUS-B1~([1,2]), to test the performance of the major color difference formulae in this region of color space using the standardized residual sum of squares (STRESS) index, correlation coefficient (r) and Spearman Rank coefficient (p) as performance metrics. The visual differences between the 65 samples and the color center were assessed by 16 observers under highly controlled viewing and illuminaiton conditions, using AATCC Gray Scale for Color Change, in three separate sittings and a total of 3120 assessments were obtained. The results showed that CIEDE2000 exhibited relatively better performance for the NCSU-B2 dataset in comparison to other equations examined. However, none of the tested equations gave a satisfactorily low STRESS value, or high correlation coefficient (r) and Spearman Rank coefficient (p) values. For the combined blue datasets (NCSU-B1 and B2), CAM02-UCS, CAM02-SCD, DIN99d and CIEDE2000 showed the best performance.