Lior Shapira and Boris Oicherman
A vast majority of color transformations applied to an image in the digital press industry are static and pre-calculated. In order to achieve the best quality on a wide variety of different images, these transformations tend to be highly conservative with respect to the use of black ink. This results in excessive use of inks, which has a negative economic and environmental impact. We present a method for dynamic computation of color transformation based on image content, with the aim to reduce ink usage. We analyze the image, and predict areas in which quality artifacts that may result from such a reduction will be masked by the image content. These areas include detailed textures, noisy areas and structure. We then replace the image CMYK values by a new combination with increased black. Our algorithm ensures negligible color shifts in the resulting image, and no visible reduction in quality. We achieve an average of over 10\% ink savings.
In Comput. Graph. Forum