Color difference evaluation for digital images

Abstract

A review is firstly presented about the history of color difference formulas and the development of color difference evaluation for digital images. This study involves two phases of psychophysical experiments. Phase I was implemented via a categorical judgment method to test the typical color difference evaluation methods for digital images, including CIELAB, CIEDE2000, CAM02-UCS, S-CIELAB, and iCAM, which indicates that iCAM performs best in the evaluation of image color difference. Two visual experiments were involved in Phase II, of which the first one was conducted to determine the threshold of image color difference and the second one was carried out based on a magnitude estimation method to evaluate the suprathreshold color difference for digital images. The visual data illustrate that the mean acceptable threshold is nearly twice the perceptible one and the different thresholds of lightness, chroma, hue and sharpness should be considered in the future modeling of color difference evaluation for digital images.