A novel error diffusion filter for digital halftoning

Error diffusion methods have been used for a long time as a method for correcting the errors resulting when representing each of the samples (pixels) of an analogue image as a single bit binary word. Several filters were previously designed and have been thoroughly analyzed. In this paper a new error diffusion filter is introduced. In this filter the weights and locations of the filter elements are inherently not deterministic. The design of this filter relies on the fact that the image signal can be considered to be of a random nature. The proposed filter produced good quality digitized images.