Abstract
An accurate printer model that is efficient enough to be used by halftoning algorithms is proposed. The signal processing model (SPM) proposed in this paper utilizes a physical model to train Adaptive Linear Combiners (ALCs), after which the average exposure of each subpixel for any input pattern can be calculated using the optimized weight vector. The SPM can be used to model multi-level halftoning and resolution enhancement, as well as traditional halftoning . The SPM is comprised of a single layer of ALC's and is adapted using the LMS algorithm. A relatively small number of training patterns suffices to obtain adequate model accuracy