Abstract

In this work we aim to identify in which manner spectral encoding and spectral processing can optimize color management and print process. In comparison with conventional color management, spectral data processing is superior in terms of reducing metamerism, has significantly higher prediction power with the same amount of training data. To prove these claims, we utilize state-of-the-art spectral and color prediction models, and we apply them on all major printing technologies in CMYK and multi-color mode. Our evaluation confirms far superior prediction power of spectral models. We also show that spectral prediction can effectively enable automation of the color management without a need for the dedicated printing session. This is only possible if print condition change is captured and modelized with the fraction of the time and measurements usually needed by conventional color management.