Predicting Camera Color Quality

Abstract

Color quality can be measured two ways. The first is target based where color-difference statistics are reported comparing image data with measurement-based colorimetric data. The second is based on measuring the camera sensor’s spectral sensitivities and calculating their similarity to a standard observer, for example, μ-factor. A computational experiment was performed where synthetic images of a variety of targets were rendered for four camera systems having μ-factors of 0.79, 0.88, 0.94, and 0.99. Each camera was profiled using the same target. Although profile color accuracy was acceptable for all the cameras, this did not predict the color accuracy for independent targets. μ-factor was a better predictor of color quality and its use is recommended when evaluating cameras for cultural heritage applications