Printer Model Integrating Genetic Algorithm for Improvement of Halftone Patterns

Abstract- The paper describes a genetic algorithm to search for halftone patterns that would print with the least visual granularity. Our approach automatically generates dithers op-timized for use in specific printing hardware. A semi-empirical model of electrophotographic printing process is applied as a fitness criterion used in evaluating performance of individual dithers. The printer model includes a function for visual contrast sensitivity and also for random print noise. Genetic searches were begun with halftone patterns of very low spa-tial frequency, very high frequency, and white noise random patterns. All of these converged to quasi-random patterns with noise power maximized just below the visual threshold. Our experiments demonstrate ability of genetic algorithms to find high fitness dithers irrespective of the printing hardware being considered. This work is part of an on-going project to develop printer calibration procedures to optimize print image quality. I