Device dependent screen optimization using evolutionary computing

摘要

Most of the halftoning algorithms are based on ideal imaging devices that can render perfect square pixels. In real printing environments this is not the case. Most imaging devices are a trade-off between the best quality and the highest speed. In this paper a screen will be designed for Agfa's newspaper-dedicated computer-to-plate imaging device Polaris. Evolutionary computing is used to search for better solutions starting from an initial population. By use of genetic algorithms the members of the population are repeatedly combined to new tiles. At each iteration step members of the current population are selected with a probability proportional to their relative fitness. Critical in the search algorithm is the fitness function. How can the computer decide if a tile is good or not? A printer model is used to generate a super sampled approximation of the result on print. Both steps of the printing process, creation of the dots on plate and the printing of plate dots on paper, are considered to build the printer model. For the plate generation a distinction is made between the fast scan mode and the slow scan mode of the imager. For the printing process a noise generation is added to model the noise caused by the structure of the paper. By use of Fourier analysis a measure of the quality of a screen tile is obtained.