Image halftoning using a visual model in error diffusion

Abstract

Continued advances in binary image printers have spurred an increased interest in the use of digital image halftoning to generate low-cost images that have the appearance of gray levels. However, at current print resolutions for desktop applications, i.e., 300–400 dots/in. (dpi), the binary noise resulting from halftoning is clearly visible at normal viewing distances. Halftoning algorithms that reduce the visibility of this noise result in smoother gray levels and higher-quality output images. We introduce a new method of gray-level image halftoning that uses visual modeling within the framework of error diffusion to improve the image quality of halftoned images.