Simplified grid message-passing algorithm with application to digital image halftoning

Based on message-passing techniques, a novel iterative grid algorithm for the general two-dimensional (2D) digital least metric (DLM) problem is proposed and applied to image halftoning. The algorithm attempts to achieve a globally optimal solution via a local-metric computation and message passing as opposed to other 2D iterative global-metric optimizations such as simulated annealing and toggle/swap scheme. A reduced-complexity version of the proposed digital image halftoning technique is demonstrated. Results show that the quality of the halftone images is comparable to that of the state-of-the-art toggle/swap algorithm. Since the algorithm is not constrained by the specific metric used, the proposed method is directly applicable to other digital image processing tasks (eg, optimal near-lossless coding or entropy-constrained halftoning).